Skip to content

I/O

Parser Package

parser

__all__ = ['read_pin', 'read_pepxml', 'extract_mzml_data'] module-attribute

extract_mzml_data(mzml_filename, scan_ids=None)

Extract scan data from an mzML file.

Parameters:

Name Type Description Default
mzml_filename str

The path to the mzML file.

 required
scan_ids list[int] or None

A list of scan IDs to extract. If None, extracts all scans.

 None

Returns:

Type Description
DataFrame

A DataFrame containing the extracted scan data with columns: - source: The source file name - scan: The scan ID - mz: The m/z values array - intensity: The intensity values array - charge: The charge state - retention_time: The retention time
Notes

This function: 1. Reads the mzML file using pyteomics 2. Extracts scan data including retention time, charge state, m/z values, and intensities 3. Filters scans based on provided scan IDs if specified 4. Returns a DataFrame with the extracted data

Source code in optimhc/parser/mzml.py 
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
def extract_mzml_data(mzml_filename, scan_ids=None):
    """
    Extract scan data from an mzML file.

    Parameters
    ----------
    mzml_filename : str
        The path to the mzML file.
    scan_ids : list[int] or None, optional
        A list of scan IDs to extract. If None, extracts all scans.

    Returns
    -------
    pd.DataFrame
        A DataFrame containing the extracted scan data with columns:
        - source: The source file name
        - scan: The scan ID
        - mz: The m/z values array
        - intensity: The intensity values array
        - charge: The charge state
        - retention_time: The retention time

    Notes
    -----
    This function:
    1. Reads the mzML file using pyteomics
    2. Extracts scan data including retention time, charge state, m/z values, and intensities
    3. Filters scans based on provided scan IDs if specified
    4. Returns a DataFrame with the extracted data
    """
    filename = mzml_filename.split("/")[-1].replace(".mzML", "")
    logger.info(f"Extracting scans from {mzml_filename}")

    scan_ids = set(scan_ids) if scan_ids is not None else None

    (
        extracted_scan_ids,
        mzml_filenames,
        intensities,
        mz_values,
        charges,
        retention_times,
    ) = ([], [], [], [], [], [])

    try:
        with mzml.read(mzml_filename) as reader:
            for spectrum in reader:
                try:
                    scan_id = int(spectrum["id"].split("scan=")[-1])

                    if scan_ids is not None and scan_id not in scan_ids:
                        continue

                    mz_array = np.array(spectrum.get("m/z array", []))
                    intensity_array = np.array(spectrum.get("intensity array", []))

                    charge = None
                    try:
                        charge = int(
                            spectrum["precursorList"]["precursor"][0]["selectedIonList"][
                                "selectedIon"
                            ][0]["charge state"]
                        )
                    except (KeyError, ValueError, IndexError):
                        pass

                    retention_time = None
                    try:
                        retention_time = float(spectrum["scanList"]["scan"][0]["scan start time"])
                    except (KeyError, ValueError, IndexError):
                        pass

                    extracted_scan_ids.append(scan_id)
                    mzml_filenames.append(filename)
                    intensities.append(intensity_array)
                    mz_values.append(mz_array)
                    charges.append(charge)
                    retention_times.append(retention_time)

                except Exception as e:
                    logger.warning(f"Skipping scan {scan_id} due to error: {e}")

    except Exception as e:
        logger.error(f"Failed to parse mzML file {mzml_filename}: {e}")
        raise RuntimeError(f"Error processing mzML file {mzml_filename}: {e}")

    data_dict = {
        "source": mzml_filenames,
        "scan": extracted_scan_ids,
        "mz": mz_values,
        "intensity": intensities,
        "charge": charges,
        "retention_time": retention_times,
    }

    scans_df = pd.DataFrame(data_dict)
    scans_df = scans_df.drop_duplicates(subset=["source", "scan"])

    logger.info(f"Successfully extracted {len(scans_df)} scans from {mzml_filename}")

    return scans_df

read_pepxml(pepxml_files, decoy_prefix='DECOY_')

Read PSMs from a list of PepXML files.

Parameters:

Name Type Description Default
pepxml_files Union[str, List[str]]

The file path to the PepXML file or a list of file paths.

 required
decoy_prefix str

The prefix used to indicate a decoy protein in the description lines of the FASTA file. Default is "DECOY_".

 'DECOY_'

Returns:

Type Description
PsmContainer

A PsmContainer object containing the PSM data.

Raises:

Type Description
ValueError

If the PepXML files were generated by Percolator or PeptideProphet.
Notes

This function: 1. Reads and parses PepXML files 2. Calculates mass difference features 3. Processes matched ions and complementary ions 4. Creates charge columns 5. Log-transforms p-values 6. Returns a PsmContainer with the processed data

Source code in optimhc/parser/pepxml.py 
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
def read_pepxml(pepxml_files, decoy_prefix="DECOY_"):
    """
    Read PSMs from a list of PepXML files.

    Parameters
    ----------
    pepxml_files : Union[str, List[str]]
        The file path to the PepXML file or a list of file paths.
    decoy_prefix : str, optional
        The prefix used to indicate a decoy protein in the description lines
        of the FASTA file. Default is "DECOY_".

    Returns
    -------
    PsmContainer
        A PsmContainer object containing the PSM data.

    Raises
    ------
    ValueError
        If the PepXML files were generated by Percolator or PeptideProphet.

    Notes
    -----
    This function:
    1. Reads and parses PepXML files
    2. Calculates mass difference features
    3. Processes matched ions and complementary ions
    4. Creates charge columns
    5. Log-transforms p-values
    6. Returns a PsmContainer with the processed data
    """
    proton = 1.00727646677
    if isinstance(pepxml_files, str):
        pepxml_files = [pepxml_files]
    psms = pd.concat([_parse_pepxml(f, decoy_prefix) for f in pepxml_files])

    # Check that these PSMs are not from Percolator or PeptideProphet:
    illegal_cols = {
        "Percolator q-Value",
        "Percolator PEP",
        "Percolator SVMScore",
    }

    if illegal_cols.intersection(set(psms.columns)):
        raise ValueError(
            "The PepXML files appear to have generated by Percolator or "
            "PeptideProphet; hence, they should not be analyzed with mokapot."
        )

    # For open modification searches:
    psms["mass_diff"] = psms["exp_mass"] - psms["calc_mass"]
    # Calculate massdiff features
    exp_mz = psms["exp_mass"] / psms["charge"] + proton
    calc_mz = psms["calc_mass"] / psms["charge"] + proton
    psms["abs_mz_diff"] = (exp_mz - calc_mz).abs()

    # Calculate matched ions and complementary ions
    if "num_matched_ions" in psms.columns and "tot_num_ions" in psms.columns:
        if (psms["tot_num_ions"] != 0).all():
            psms["matched_ions_ratio"] = psms["num_matched_ions"] / psms["tot_num_ions"]

    # Log number of candidates:
    if "num_matched_peptides" in psms.columns:
        psms["num_matched_peptides"] = np.log10(psms["num_matched_peptides"])

    # Create charge columns:
    psms = pd.concat([psms, pd.get_dummies(psms["charge"], prefix="charge")], axis=1)

    # psms = psms.drop("charge", axis=1)
    # -log10 p-values
    nonfeat_cols = [
        "ms_data_file",
        "scan",
        "spectrum",
        "label",
        "calc_mass",  # Retain calc_mass for FlashLFQ compatibility
        "peptide",
        "proteins",
        "charge",
        "retention_time",
    ]

    feat_cols = [c for c in psms.columns if c not in nonfeat_cols]
    psms = psms.apply(_log_features, features=feat_cols)
    rescoring_features = {"Original": feat_cols}

    return PsmContainer(
        psms=psms,
        label_column="label",
        scan_column="scan",
        spectrum_column="spectrum",
        ms_data_file_column="ms_data_file",
        peptide_column="peptide",
        protein_column="proteins",
        charge_column="charge",
        rescoring_features=rescoring_features,
        hit_rank_column="rank",
        retention_time_column="retention_time",
        calculated_mass_column="calc_mass",
    )

read_pin(pin_files, retention_time_column=None, remove_pre_nxt_aa=False)

Read PSMs from a Percolator INput (PIN) file.

Parameters:

Name Type Description Default
pin_files Union[str, List[str]]

The file path to the PIN file or a list of file paths.

 required
retention_time_column Optional[str]

The column containing the retention time. If None, no retention time will be included.

 None

Returns:

Type Description
PsmContainer

A PsmContainer object containing the PSM data.
Notes

This function: 1. Reads PIN file(s) into a DataFrame 2. Identifies required columns (case-insensitive) 3. Processes scan IDs and hit ranks (Only support FragPipe PIN) 4. Converts data types appropriately 5. Creates a PsmContainer with the processed data

Source code in optimhc/parser/pin.py 
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
def read_pin(
    pin_files: Union[str, List[str]],
    retention_time_column: Optional[str] = None,
    remove_pre_nxt_aa: bool = False,
) -> PsmContainer:
    """
    Read PSMs from a Percolator INput (PIN) file.

    Parameters
    ----------
    pin_files : Union[str, List[str]]
        The file path to the PIN file or a list of file paths.
    retention_time_column : Optional[str], optional
        The column containing the retention time. If None, no retention time
        will be included.

    Returns
    -------
    PsmContainer
        A PsmContainer object containing the PSM data.

    Notes
    -----
    This function:
    1. Reads PIN file(s) into a DataFrame
    2. Identifies required columns (case-insensitive)
    3. Processes scan IDs and hit ranks (Only support FragPipe PIN)
    4. Converts data types appropriately
    5. Creates a PsmContainer with the processed data
    """
    logger.info("Reading PIN file(s) into PsmContainer.")
    if isinstance(pin_files, str):
        pin_files = [pin_files]

    pin_df = pd.concat([_read_single_pin_as_df(pin_file) for pin_file in pin_files])
    logger.info(f"Read {len(pin_df)} PSMs from {len(pin_files)} PIN files.")
    logger.debug(pin_df.head())
    logger.debug(pin_df.columns)
    logger.debug(pin_df.iloc[0])

    def find_required_columns(col: str, columns: List[str]) -> str:
        """
        Case-insensitive search for a column in the DataFrame.
        Returns the matching column name with original casing.
        """
        col_lower = col.lower()
        column_map = {c.lower(): c for c in columns}
        if col_lower not in column_map:
            raise ValueError(f"Column '{col}' not found in PSM data (case-insensitive).")
        return column_map[col_lower]

    # non-feature columns (case-insensitive search)
    label = find_required_columns("Label", pin_df.columns)
    scan = find_required_columns("ScanNr", pin_df.columns)
    specid = find_required_columns("SpecId", pin_df.columns)
    peptide = find_required_columns("Peptide", pin_df.columns)
    protein = find_required_columns("Proteins", pin_df.columns)

    # Comet: P2PI20160713_pilling_C1RA2_BB72_P1_31_3_1
    # Fragpipe: P2PI20160713_pilling_C1RA2_BB72_P1.3104.3104.2_1

    # Try to parse rank from SpecId
    def parse_specid(specid: str) -> Tuple[str, int]:
        if "_" in specid:
            parts = specid.rsplit("_", 1)
            if len(parts) != 2:
                logger.warning(f"SpecId format unexpected: {specid}, using default rank 1")
                return 1
            try:
                hit_rank = int(parts[1])
                return hit_rank
            except ValueError:
                logger.warning(f"Could not parse rank from SpecId: {specid}, using default rank 1")
                return 1
        else:
            return 1

    hit_rank = "rank"
    if "rank" in [c.lower() for c in pin_df.columns]:
        pass
    else:
        # Parse SpecId to extract hit rank and update both columns
        pin_df["rank"] = pin_df[specid].apply(parse_specid)

    retention_time_column = (
        find_required_columns(retention_time_column, pin_df.columns)
        if retention_time_column
        else None
    )

    # col: charge_[1,2,3,...] = 0, 1
    charge_map = {
        col: int(re.search(r"(\d+)", col).group(1))
        for col in pin_df.columns
        if re.search(r"charge[_]?(\d+)", col, re.IGNORECASE)
    }

    def extract_charge(row):
        for col, num in charge_map.items():
            if int(float(row[col])) == 1:
                return num
        return None

    pin_df["Charge"] = pin_df.apply(extract_charge, axis=1)

    # feature columns: columns that are not non-feature columns
    non_feature_columns = [label, scan, specid, peptide, protein, hit_rank, "Charge"]
    feature_columns = [col for col in pin_df.columns if col not in non_feature_columns]

    logger.info(
        f"Columns: label={label}, scan={scan}, specid={specid}, peptide={peptide}, "
        f"protein={protein}, hit_rank={hit_rank}, retention_time={retention_time_column}, "
        f"features={feature_columns}"
    )

    pin_df[scan] = pin_df[scan].astype(int)
    pin_df[specid] = pin_df[specid].astype(str)
    pin_df[peptide] = pin_df[peptide].astype(str)
    pin_df[protein] = pin_df[protein].astype(str)
    pin_df[hit_rank] = pin_df[hit_rank].astype(float).astype(int)
    pin_df["Charge"] = pin_df["Charge"].astype(float).astype(int)
    if retention_time_column:
        pin_df[retention_time_column] = pin_df[retention_time_column].astype(float)
    for col in feature_columns:
        pin_df[col] = pin_df[col].astype(float)

    # label = 1 for target, -1 for decoy. Convert to Boolean.
    pin_df[label] = pin_df[label] == "1"
    rescoring_features = {"Original": feature_columns}

    return PsmContainer(
        psms=pin_df,
        label_column=label,
        scan_column=scan,
        spectrum_column=specid,
        ms_data_file_column=None,
        peptide_column=peptide,
        protein_column=protein,
        charge_column="Charge",
        rescoring_features=rescoring_features,
        hit_rank_column=hit_rank,
        retention_time_column=retention_time_column,
    )

PepXML Parser

pepxml

logger = logging.getLogger(__name__) module-attribute

PsmContainer(psms, label_column, scan_column, spectrum_column, ms_data_file_column, peptide_column, protein_column, rescoring_features, hit_rank_column=None, charge_column=None, retention_time_column=None, calculated_mass_column=None, metadata_column=None)

A container for managing peptide-spectrum matches (PSMs) in immunopeptidomics rescoring pipelines.

Parameters:

Name Type Description Default
psms DataFrame

DataFrame containing the PSM data.

 required
label_column str

Column containing the label (True for target, False for decoy).

 required
scan_column str

Column containing the scan number.

 required
spectrum_column str

Column containing the spectrum identifier.

 required
ms_data_file_column str

Column containing the MS data file that the PSM originated from.

 required
peptide_column str

Column containing the peptide sequence.

 required
protein_column str

Column containing the protein accessions.

 required
rescoring_features dict of str to list of str

Dictionary of feature columns for rescoring.

 required
hit_rank_column str

Column containing the hit rank.

 None
charge_column str

Column containing the charge state.

 None
retention_time_column str

Column containing the retention time.

 None
calculated_mass_column str

Column containing the calculated mass.

 None
metadata_column str

Column containing metadata.

 None

Attributes:

Name Type Description
psms DataFrame

Copy of the DataFrame containing the PSM data.
target_psms DataFrame

DataFrame containing only target PSMs (label = True).
decoy_psms DataFrame

DataFrame containing only decoy PSMs (label = False).
peptides list of str

List containing all peptides from the PSM data.
columns list of str

List of column names in the PSM DataFrame.
rescoring_features dict of str to list of str

Dictionary of rescoring feature columns in the PSM DataFrame.
Source code in optimhc/psm_container.py 
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
def __init__(
    self,
    psms: pd.DataFrame,
    label_column: str,
    scan_column: str,
    spectrum_column: str,
    ms_data_file_column: str,
    peptide_column: str,
    protein_column: str,
    rescoring_features: Dict[str, List[str]],
    hit_rank_column: Optional[str] = None,
    charge_column: Optional[str] = None,
    retention_time_column: Optional[str] = None,
    calculated_mass_column: Optional[str] = None,
    metadata_column: Optional[str] = None,
):
    self._psms = psms.copy()
    self._psms.reset_index(drop=True, inplace=True)
    self.label_column = label_column
    self.scan_column = scan_column
    self.spectrum_column = spectrum_column
    self.ms_data_file_column = ms_data_file_column
    self.peptide_column = peptide_column
    self.protein_column = protein_column
    self.hit_rank_column = hit_rank_column
    self.retention_time_column = retention_time_column
    self.metadata_column = metadata_column
    self.rescoring_features = rescoring_features
    self.charge_column = charge_column
    self.calculated_mass_column = calculated_mass_column
    # rescore result column
    self.rescore_result_column = None

    # check if the columns are in the dataframe
    def check_column(col):
        if col and col not in psms.columns:
            raise ValueError(f"Column '{col}' not found in PSM data.")

    check_column(label_column)
    check_column(scan_column)
    check_column(spectrum_column)
    check_column(ms_data_file_column)
    check_column(peptide_column)
    check_column(protein_column)
    check_column(hit_rank_column)
    check_column(retention_time_column)
    check_column(charge_column)
    check_column(calculated_mass_column)

    # ensure the label column is boolean
    if psms[label_column].dtype != "bool":
        raise ValueError(f"Column '{label_column}' must be boolean.")

    if psms[label_column].nunique() == 1 and psms[label_column].iloc[0]:
        raise ValueError("All PSMs are labeled as target. No decoy PSMs found.")
    elif psms[label_column].nunique() == 1 and not psms[label_column].iloc[0]:
        raise ValueError("All PSMs are labeled as decoy. No target PSMs found.")

    def check_metadata_column(col):
        # check the type is Dict[str, Dict[str, str]]
        if col and col not in psms.columns:
            raise ValueError(f"Column '{col}' not found in PSM data.")
        if not all(isinstance(x, dict) for x in self._psms[col]):
            raise ValueError(f"Column '{col}' must contain dictionaries.")

    if metadata_column:
        check_metadata_column(metadata_column)

    def check_rescoring_features(features: Dict[str, List[str]]):
        for key, cols in features.items():
            for col in cols:
                if col not in psms.columns:
                    raise ValueError(
                        f"Column '{col}' not found in PSM data for feature '{key}'."
                    )

    check_rescoring_features(rescoring_features)

    # check if the number of decoy psms is not 0
    if len(self.decoy_psms) == 0:
        logger.error("No decoy PSMs found. Please check the decoy prefix.")
        raise ValueError("No decoy PSMs found.")

    logger.info("PsmContainer initialized with %d PSM entries.", len(self._psms))
    if self.ms_data_file_column:
        logger.info(
            "PSMs originated from %d MS data file(s).",
            len(self._psms[ms_data_file_column].unique()),
        )
    logger.info("target psms: %d", len(self.target_psms))
    logger.info("decoy psms: %d", len(self.decoy_psms))
    logger.info("unique peptides: %d", len(np.unique(self.peptides)))
    logger.info("rescoring features: %s", rescoring_features)

psms property

Get a copy of the PSM DataFrame to prevent external modification.

Returns:

Type Description
DataFrame

A copy of the PSM DataFrame.

target_psms property

Get a DataFrame containing only target PSMs.

Returns:

Type Description
DataFrame

DataFrame with only target PSMs (label = True).

decoy_psms property

Get a DataFrame containing only decoy PSMs.

Returns:

Type Description
DataFrame

DataFrame with only decoy PSMs (label = False).

columns property

Get the column names of the PSM DataFrame.

Returns:

Type Description
list of str

List of column names.

feature_columns property

Get a list of all feature columns in the PSM DataFrame.

Returns:

Type Description
list of str

List of feature column names.

feature_sources property

Get a list of all feature sources in the PSM DataFrame.

Returns:

Type Description
list of str

List of feature source names.

peptides property

Get the peptide sequences from the PSM data.

Returns:

Type Description
list of str

List of peptide sequences.

ms_data_files property

Get the MS data files from the PSM data.

Returns:

Type Description
list of str

List of MS data file names.

scan_ids property

Get the scan numbers from the PSM data.

Returns:

Type Description
list of int

List of scan numbers.

charges property

Get the charge states from the PSM data.

Returns:

Type Description
list of int

List of charge states.

metadata property

Get the metadata from the PSM data.

Returns:

Type Description
Series

Series containing metadata for each PSM.

spectrum_ids property

Get the spectrum identifiers from the PSM data.

Returns:

Type Description
list of str

List of spectrum identifiers.

identifier_columns property

Get the columns that uniquely identify each PSM.

Returns:

Type Description
list of str

List of identifier column names.

__len__()

Get the number of PSMs in the container.

Returns:

Type Description
int

Number of PSMs.
Source code in optimhc/psm_container.py 
168
169
170
171
172
173
174
175
176
177
def __len__(self) -> int:
    """
    Get the number of PSMs in the container.

    Returns
    -------
    int
        Number of PSMs.
    """
    return len(self._psms)

copy()

Return a deep copy of the PsmContainer object.

Returns:

Type Description
PsmContainer

A deep copy of the current PsmContainer.
Source code in optimhc/psm_container.py 
328
329
330
331
332
333
334
335
336
337
338
339
def copy(self) -> "PsmContainer":
    """
    Return a deep copy of the PsmContainer object.

    Returns
    -------
    PsmContainer
        A deep copy of the current PsmContainer.
    """
    import copy

    return copy.deepcopy(self)

__repr__()

Return a string representation of the PsmContainer.

Returns:

Type Description
str

String summary of the PsmContainer.
Source code in optimhc/psm_container.py 
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
def __repr__(self) -> str:
    """
    Return a string representation of the PsmContainer.

    Returns
    -------
    str
        String summary of the PsmContainer.
    """
    return (
        f"PsmContainer with {len(self)} PSMs\n"
        f"\t - Target PSMs: {len(self.target_psms)}\n"
        f"\t - Decoy PSMs: {len(self.decoy_psms)}\n"
        f"\t - Unique Peptides: {len(np.unique(self.peptides))}\n"
        f"\t - Unique Spectra: {len(self._psms[self.spectrum_column].unique())}\n"
        f"\t - Rescoring Features: {self.rescoring_features}\n"
    )

drop_features(features)

Drop specified features from the PSM DataFrame.

Parameters:

Name Type Description Default
features list of str

List of feature column names to drop.

 required

Raises:

Type Description
ValueError

If any of the features do not exist in the DataFrame.
Source code in optimhc/psm_container.py 
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
def drop_features(self, features: List[str]) -> None:
    """
    Drop specified features from the PSM DataFrame.

    Parameters
    ----------
    features : list of str
        List of feature column names to drop.

    Raises
    ------
    ValueError
        If any of the features do not exist in the DataFrame.
    """
    missing_features = [f for f in features if f not in self._psms.columns]
    if missing_features:
        raise ValueError(f"Features not found in PSM data: {missing_features}")

    self._psms.drop(columns=features, inplace=True)
    # Create a list of sources to update
    sources_to_update = []
    for source, cols in self.rescoring_features.items():
        self.rescoring_features[source] = [col for col in cols if col not in features]
        if not self.rescoring_features[source]:
            sources_to_update.append(source)

    logger.info(
        f"Sources to be removed: {sources_to_update}. Because all the features are removed."
    )
    # Remove sources with no features left
    for source in sources_to_update:
        del self.rescoring_features[source]

drop_source(source)

Drop all features associated with a specific source from the PSM DataFrame.

Parameters:

Name Type Description Default
source str

Name of the source to drop.

 required

Raises:

Type Description
ValueError

If the source does not exist in the rescoring features.
Source code in optimhc/psm_container.py 
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
def drop_source(self, source: str) -> None:
    """
    Drop all features associated with a specific source from the PSM DataFrame.

    Parameters
    ----------
    source : str
        Name of the source to drop.

    Raises
    ------
    ValueError
        If the source does not exist in the rescoring features.
    """
    if source not in self.rescoring_features:
        raise ValueError(f"Source '{source}' not found in rescoring features.")
    self.drop_features(self.rescoring_features[source])

add_metadata(metadata_df, psms_key, metadata_key, source)

Merge new metadata into the PSM DataFrame based on specified columns. Metadata from the specified source is stored as a nested dictionary inside the metadata column.

Parameters:

Name Type Description Default
metadata_df DataFrame

DataFrame containing new metadata to add.

 required
psms_key str or list of str

Column name(s) in the PSM data to merge on.

 required
metadata_key str or list of str

Column name(s) in the metadata data to merge on.

 required
source str

Name of the source of the new metadata.

 required
Source code in optimhc/psm_container.py 
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
def add_metadata(
    self,
    metadata_df: pd.DataFrame,
    psms_key: Union[str, List[str]],
    metadata_key: Union[str, List[str]],
    source,
) -> None:
    """
    Merge new metadata into the PSM DataFrame based on specified columns.
    Metadata from the specified source is stored as a nested dictionary inside the metadata column.

    Parameters
    ----------
    metadata_df : pd.DataFrame
        DataFrame containing new metadata to add.
    psms_key : str or list of str
        Column name(s) in the PSM data to merge on.
    metadata_key : str or list of str
        Column name(s) in the metadata data to merge on.
    source : str
        Name of the source of the new metadata.
    """
    if self.metadata_column is None:
        logger.info("No existing metadata column. Creating new metadata column.")
        self.metadata_column = "metadata"
        self._psms["metadata"] = [{} for _ in range(len(self._psms))]

    metadata_cols = [col for col in metadata_df.columns if col not in metadata_key]
    merged_df = self.psms.merge(
        metadata_df, left_on=psms_key, right_on=metadata_key, how="left"
    )
    if source in self._psms["metadata"]:
        logger.warning(f"{source} already exists in metadata. Overwriting.")
    for col in metadata_cols:
        merged_df["metadata"] = merged_df.apply(
            lambda row: {
                **row["metadata"],
                source: (
                    {col: row[col]}
                    if source not in row["metadata"]
                    else {**row["metadata"][source], col: row[col]}
                ),
            },
            axis=1,
        )

    self._psms["metadata"] = merged_df["metadata"]

get_top_hits(n=1)

Get the top n hits based on the hit rank column. If the hit rank column is not specified, returns the original PSMs.

Parameters:

Name Type Description Default
n int

The number of top hits to return. Default is 1.

 1

Returns:

Type Description
PsmContainer

A new PsmContainer object containing the top n hits.
Source code in optimhc/psm_container.py 
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
def get_top_hits(self, n: int = 1):
    """
    Get the top n hits based on the hit rank column.
    If the hit rank column is not specified, returns the original PSMs.

    Parameters
    ----------
    n : int, optional
        The number of top hits to return. Default is 1.

    Returns
    -------
    PsmContainer
        A new PsmContainer object containing the top n hits.
    """
    if self.hit_rank_column is None:
        logger.warning("Rank column not specified. Return the original PSMs.")
        return self.copy()

    psms = self.copy()
    psms._psms = psms._psms[psms._psms[self.hit_rank_column] <= n]
    return psms

add_features(features_df, psms_key, feature_key, source, suffix=None)

Merge new features into the PSM DataFrame based on specified columns.

This method performs a left join between the PSM data and feature data, ensuring that all PSMs are preserved while adding new features. It handles column name conflicts through optional suffixing and maintains feature source tracking.

Parameters:

Name Type Description Default
features_df DataFrame

DataFrame containing new features to add.

 required
psms_key str or list of str

Column name(s) in the PSM data to merge on.

 required
feature_key str or list of str

Column name(s) in the features data to merge on.

 required
source str

Name of the source of the new features (e.g., 'deeplc', 'netmhc').

 required
suffix str

Suffix to add to the new columns if there's a name conflict. Required when new feature columns have the same names as existing columns. For example, if adding features from different sources (e.g., 'score' from DeepLC and NetMHC), use suffixes like '_deeplc' or '_netmhc' to distinguish them.

 None

Returns:

Type Description
None

Raises:

Type Description
ValueError

If duplicate columns exist without suffix. If merging features changes the number of PSMs.
Notes

The method follows these steps: 1. Validates input and prepares merge keys 2. Checks for column name conflicts 3. Manages feature source: if the source already exists, it will be overwritten 4. Performs left join merge 5. Verifies data integrity

Suffix Usage

The suffix parameter is used to handle column name conflicts: - When adding features from different sources that might have the same column names - When you want to keep both the original and new features with the same name - When you need to track the source of features in the column names

If suffix is not provided and there are duplicate column names: - The method will raise a ValueError - You must either provide a suffix or rename the columns before adding

Examples:

>>> container = PsmContainer(...)
>>> # Adding features without suffix (no conflicts)
>>> features_df1 = pd.DataFrame({
...     'scan': [1, 2, 3],
...     'feature1': [0.1, 0.2, 0.3],
...     'feature2': [0.4, 0.5, 0.6]
... })
>>> container.add_features(
...     features_df1,
...     psms_key='scan',
...     feature_key='scan',
...     source='source1'
... )
>>> # Adding features with suffix (handling conflicts)
>>> features_df2 = pd.DataFrame({
...     'scan': [1, 2, 3],
...     'score': [0.8, 0.9, 0.7],  # This would conflict with existing 'score'
...     'feature3': [0.7, 0.8, 0.9]
... })
>>> container.add_features(
...     features_df2,
...     psms_key='scan',
...     feature_key='scan',
...     source='source2',
...     suffix='_new'  # 'score' becomes 'score_new'
... )
Source code in optimhc/psm_container.py 
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
def add_features(
    self,
    features_df: pd.DataFrame,
    psms_key: Union[str, List[str]],
    feature_key: Union[str, List[str]],
    source: str,
    suffix: Optional[str] = None,
) -> None:
    """Merge new features into the PSM DataFrame based on specified columns.

    This method performs a left join between the PSM data and feature data,
    ensuring that all PSMs are preserved while adding new features. It handles
    column name conflicts through optional suffixing and maintains feature source
    tracking.

    Parameters
    ----------
    features_df : pd.DataFrame
        DataFrame containing new features to add.
    psms_key : str or list of str
        Column name(s) in the PSM data to merge on.
    feature_key : str or list of str
        Column name(s) in the features data to merge on.
    source : str
        Name of the source of the new features (e.g., 'deeplc', 'netmhc').
    suffix : str, optional
        Suffix to add to the new columns if there's a name conflict.
        Required when new feature columns have the same names as existing columns.
        For example, if adding features from different sources (e.g., 'score' from
        DeepLC and NetMHC), use suffixes like '_deeplc' or '_netmhc' to distinguish them.

    Returns
    -------
    None

    Raises
    ------
    ValueError
        If duplicate columns exist without suffix.
        If merging features changes the number of PSMs.

    Notes
    -----
    The method follows these steps:
    1. Validates input and prepares merge keys
    2. Checks for column name conflicts
    3. Manages feature source: if the source already exists, it will be overwritten
    4. Performs left join merge
    5. Verifies data integrity

    Suffix Usage
    -----------
    The suffix parameter is used to handle column name conflicts:
    - When adding features from different sources that might have the same column names
    - When you want to keep both the original and new features with the same name
    - When you need to track the source of features in the column names

    If suffix is not provided and there are duplicate column names:
    - The method will raise a ValueError
    - You must either provide a suffix or rename the columns before adding

    Examples
    --------
    >>> container = PsmContainer(...)
    >>> # Adding features without suffix (no conflicts)
    >>> features_df1 = pd.DataFrame({
    ...     'scan': [1, 2, 3],
    ...     'feature1': [0.1, 0.2, 0.3],
    ...     'feature2': [0.4, 0.5, 0.6]
    ... })
    >>> container.add_features(
    ...     features_df1,
    ...     psms_key='scan',
    ...     feature_key='scan',
    ...     source='source1'
    ... )
    >>> # Adding features with suffix (handling conflicts)
    >>> features_df2 = pd.DataFrame({
    ...     'scan': [1, 2, 3],
    ...     'score': [0.8, 0.9, 0.7],  # This would conflict with existing 'score'
    ...     'feature3': [0.7, 0.8, 0.9]
    ... })
    >>> container.add_features(
    ...     features_df2,
    ...     psms_key='scan',
    ...     feature_key='scan',
    ...     source='source2',
    ...     suffix='_new'  # 'score' becomes 'score_new'
    ... )
    """
    if isinstance(psms_key, str):
        psms_key = [psms_key]

    if isinstance(feature_key, str):
        feature_key = [feature_key]

    new_feature_cols = [col for col in features_df.columns if col not in feature_key]

    for cols in new_feature_cols:
        if cols in self._psms.columns:
            logger.warning(f"Column '{cols}' already exists in PSM data.")
            if suffix is None:
                logger.warning("No suffix provided. Using default suffix ")
                raise ValueError("Duplicate columns exist. No suffix provided.")
            else:
                logger.warning(f"Suffix '{suffix}' provided. Using suffix '{suffix}'.")
    logger.info(f"Adding {len(new_feature_cols)} new features from {source}.")

    if not new_feature_cols:
        logger.warning("No new features to add. Check the feature key and PSMs key.")
        logger.warning(f"Feature key: {feature_key}; PSMs key: {psms_key}")

    if source in self.rescoring_features:
        logger.warning(f"{source} already exists in rescoring features. Overwriting.")
        self.drop_source(source)

    # TODO: reluctant logic
    if suffix is None:
        suffixes = ("", "")
    else:
        suffixes = ("", suffix)

    self.rescoring_features[source] = [col + suffixes[1] for col in new_feature_cols]
    features_df = features_df.rename(
        columns={col: col + suffixes[1] for col in new_feature_cols}
    )
    original_len = len(self._psms)
    # avoid merge the right key to the psms
    self._psms = self._psms.merge(
        features_df, left_on=psms_key, right_on=feature_key, how="left"
    )

    if feature_key != psms_key:
        cols_to_drop = [
            col for col in feature_key if col not in psms_key and col in self._psms.columns
        ]
        if cols_to_drop:
            logger.debug(f"Dropping columns from feature_key not in psms_key: {cols_to_drop}")
            self._psms.drop(columns=cols_to_drop, inplace=True)

    if len(self._psms) != original_len:
        raise ValueError(
            "Merging features resulted in a change in the number of PSMs. Check for duplicate keys."
        )

add_features_by_index(features_df, source, suffix=None)

Merge new features into the PSM DataFrame based on the DataFrame index.

Parameters:

Name Type Description Default
features_df DataFrame

DataFrame containing new features to add.

 required
source str

Name of the source of the new features.

 required
suffix str

Suffix to add to the new columns if there's a name conflict.

 None
Source code in optimhc/psm_container.py 
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
def add_features_by_index(
    self, features_df: pd.DataFrame, source: str, suffix: Optional[str] = None
) -> None:
    """
    Merge new features into the PSM DataFrame based on the DataFrame index.

    Parameters
    ----------
    features_df : pd.DataFrame
        DataFrame containing new features to add.
    source : str
        Name of the source of the new features.
    suffix : str, optional
        Suffix to add to the new columns if there's a name conflict.
    """
    new_feature_cols = [col for col in features_df.columns]
    for col in new_feature_cols:
        if col in self._psms.columns:
            logger.warning(f"Column '{col}' already exists in PSM data.")
            if suffix is None:
                logger.warning("No suffix provided. Using default suffix.")
                raise ValueError("Duplicate columns exist. No suffix provided.")
            else:
                logger.warning(f"Suffix '{suffix}' provided. Using suffix '{suffix}'.")

    logger.info(f"Adding {len(new_feature_cols)} new features from {source} by index.")

    if not new_feature_cols:
        logger.warning("No new features to add.")
        raise ValueError("No new features to add.")

    if source in self.rescoring_features:
        logger.warning(f"{source} already exists in rescoring features. Overwriting.")
        self.drop_source(source)

    if suffix is None:
        suffixes = ("", "")
    else:
        suffixes = ("", suffix)

    self.rescoring_features[source] = [col + suffixes[1] for col in new_feature_cols]
    features_df.rename(
        columns={col: col + suffixes[1] for col in new_feature_cols}, inplace=True
    )
    original_len = len(self._psms)
    self._psms = self._psms.merge(
        features_df,
        left_index=True,
        right_index=True,
        how="left",  # Perform a left join to preserve all original PSM data
    )

    # Ensure that the merge did not change the number of rows in the PSM DataFrame
    if len(self._psms) != original_len:
        raise ValueError(
            "Merging features resulted in a change in the number of PSMs. Check for duplicate indices."
        )

add_results(results_df, psms_key, result_key)

Add results of rescore engine to the PSM DataFrame based on specified columns.

Parameters:

Name Type Description Default
results_df DataFrame

DataFrame containing new results to add.

 required
psms_key str or list of str

Column name(s) in the PSM data to merge on.

 required
result_key str or list of str

Column name(s) in the results data to merge on.

 required
Source code in optimhc/psm_container.py 
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
def add_results(
    self,
    results_df: pd.DataFrame,
    psms_key: Union[str, List[str]],
    result_key: Union[str, List[str]],
) -> None:
    """
    Add results of rescore engine to the PSM DataFrame based on specified columns.

    Parameters
    ----------
    results_df : pd.DataFrame
        DataFrame containing new results to add.
    psms_key : str or list of str
        Column name(s) in the PSM data to merge on.
    result_key : str or list of str
        Column name(s) in the results data to merge on.
    """
    if self.rescore_result_column is not None:
        logger.warning("Rescore result column already exists. Overwriting.")

    if set(self._psms.columns) & set(results_df.columns):
        raise ValueError(
            "Duplicate columns exist. Please rename the columns in the results data."
        )

    self.rescore_result_column = result_key
    self._psms = self._psms.merge(
        results_df,
        left_on=psms_key,
        right_on=result_key,
        how="left",
        validate="one_to_one",
    )
    self._psms.drop(columns=result_key, inplace=True)
    logger.info("Added rescore results to PSM data.")

write_pin(output_file, style='default', source=None)

Write the PSM data to a Percolator PIN file, supporting both generic Percolator and MSBooster-compatible formats. The style parameter is actually used to output a unified pin format file to benchmark the performance of different rescoring methods.

Parameters:

Name Type Description Default
output_file str

Path to the output PIN file.

 required
style str

If set to 'msbooster', outputs only the columns required by MSBooster (SpecId, Label, ScanNr, retentiontime, rank, hyperscore or log10_evalue, Peptide, Proteins). If set to 'default', outputs all features specified in rescoring_features, plus required Percolator columns.

 'default'
source list of str

List of feature sources to include. If None, includes all sources.

 None

Returns:

Type Description
DataFrame

The DataFrame written to the PIN file.
Notes
  • The first three columns are always: SpecID, Label, ScanNr.
  • For 'msbooster' style, the columns are: SpecId, Label, ScanNr, retentiontime, rank, hyperscore or log10_evalue, Peptide, Proteins.
  • If hit_rank_column is not specified, rank is set to 1 for all rows.
  • Either 'hyperscore' or 'expect' must be present in features; for 'expect', the column is written as 'log10_evalue'.
  • The 'log10_evalue' column should contain the base-10 logarithm of the e-value.
  • The 'Peptide' column is formatted with underscores (e.g., _.PEPTIDE._).
  • For standard format, all features from rescoring_features are appended between ScanNr and Peptide columns.
  • The 'Proteins' column is a semicolon-separated list if stored as a list or tuple.
  • Label column is converted to 1 (target) and -1 (decoy), as required by Percolator.

Example output (default style): SpecId Label ScanNr feature1 feature2 ... Peptide Proteins

Example output (msbooster style): SpecId Label ScanNr retentiontime rank hyperscore Peptide Proteins or SpecId Label ScanNr retentiontime rank log10_evalue Peptide Proteins

Raises:

Type Description
ValueError

If required columns are missing for the selected style.
Source code in optimhc/psm_container.py 
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
def write_pin(
    self, output_file: str, style: str = "default", source: List[str] = None
) -> None:
    """
    Write the PSM data to a Percolator PIN file, supporting both generic Percolator and MSBooster-compatible formats.
    The style parameter is actually used to output a unified pin format file to benchmark the performance of different rescoring methods.

    Parameters
    ----------
    output_file : str
        Path to the output PIN file.
    style : str, optional
        If set to 'msbooster', outputs only the columns required by MSBooster (SpecId, Label, ScanNr, retentiontime, rank, hyperscore or log10_evalue, Peptide, Proteins).
        If set to 'default', outputs all features specified in `rescoring_features`, plus required Percolator columns.
    source : list of str, optional
        List of feature sources to include. If None, includes all sources.

    Returns
    -------
    pd.DataFrame
        The DataFrame written to the PIN file.

    Notes
    -----
    - The first three columns are always: SpecID, Label, ScanNr.
    - For 'msbooster' style, the columns are: SpecId, Label, ScanNr, retentiontime, rank, hyperscore or log10_evalue, Peptide, Proteins.
    - If `hit_rank_column` is not specified, rank is set to 1 for all rows.
    - Either 'hyperscore' or 'expect' must be present in features; for 'expect', the column is written as 'log10_evalue'.
    - The 'log10_evalue' column should contain the base-10 logarithm of the e-value.
    - The 'Peptide' column is formatted with underscores (e.g., `_.PEPTIDE._`).
    - For standard format, all features from `rescoring_features` are appended between ScanNr and Peptide columns.
    - The 'Proteins' column is a semicolon-separated list if stored as a list or tuple.
    - Label column is converted to 1 (target) and -1 (decoy), as required by Percolator.

    Example output (default style):
        SpecId	Label	ScanNr	feature1	feature2	...	Peptide	Proteins

    Example output (msbooster style):
        SpecId	Label	ScanNr	retentiontime	rank	hyperscore	Peptide	Proteins
        or
        SpecId	Label	ScanNr	retentiontime	rank	log10_evalue	Peptide	Proteins

    Raises
    ------
    ValueError
        If required columns are missing for the selected style.
    """
    df = self._psms.copy()
    # Check if the label column is str
    # Case1: label column is str
    if df[self.label_column].dtype == "str":
        df["PercolatorLabel"] = df[self.label_column].map({"True": 1, "False": -1})
    # Case2: label column is bool
    elif df[self.label_column].dtype == "bool":
        df["PercolatorLabel"] = df[self.label_column].map({True: 1, False: -1})
    else:
        # try to convert to bool
        logger.warning("Label column is not str or bool. Converting to bool.")
        df["PercolatorLabel"] = df[self.label_column].astype(bool).map({True: 1, False: -1})
    logger.info("Writing PIN file to %s", output_file)
    logger.info("Using style: %s", style)

    feature_cols = []
    if source is None:
        for _, cols in self.rescoring_features.items():
            feature_cols.extend(cols)
    else:
        for s in source:
            if s not in self.rescoring_features:
                raise ValueError(f"Source '{s}' not found in rescoring features.")
            feature_cols.extend(self.rescoring_features[s])

    pin_df = pd.DataFrame()
    pin_df["SpecId"] = df[self.spectrum_column]
    pin_df["Label"] = df["PercolatorLabel"]
    pin_df["ScanNr"] = df[self.scan_column]

    if style == "msbooster":
        if self.retention_time_column:
            pin_df["retentiontime"] = df[self.retention_time_column]
        else:
            raise ValueError("Retention time column is required for msbooster style.")

        pin_df["rank"] = df[self.hit_rank_column].astype(int) if self.hit_rank_column else 1
        if "hyperscore" in self.feature_columns:
            pin_df["hyperscore"] = df["hyperscore"]
        elif "expect" in self.feature_columns:
            pin_df["log10_evalue"] = df["expect"]
        else:
            raise ValueError(
                "Either 'hyperscore' or 'expect' column is required for msbooster style."
            )

        # Add other features and jump the hyperscore or expect column
        for col in feature_cols:
            if col not in [
                "hyperscore",
                "expect",
                self.hit_rank_column,
                self.retention_time_column,
            ]:
                pin_df[col] = df[col]

        # PEPTIDE -> _.PEPTIDE._
        # Add _. at the front and ._ at the end of the peptide column
        pin_df["Peptide"] = df[self.peptide_column].apply(
            lambda x: f"_.{x}._" if isinstance(x, str) else x
        )

    elif style == "default":
        for col in feature_cols:
            pin_df[col] = df[col]
        pin_df["Peptide"] = df[self.peptide_column]
    else:
        raise ValueError(f"Unknown style: {style}. Use 'msbooster' or 'default'.")

    pin_df["Proteins"] = df[self.protein_column].apply(
        lambda x: ";".join(x) if isinstance(x, (list, tuple)) else x
    )
    pin_df = self._convert_float_to_int(pin_df)
    pin_df.to_csv(output_file, sep="\t", index=False)
    logger.info("PIN file written to %s", output_file)
    return pin_df

read_pepxml(pepxml_files, decoy_prefix='DECOY_')

Read PSMs from a list of PepXML files.

Parameters:

Name Type Description Default
pepxml_files Union[str, List[str]]

The file path to the PepXML file or a list of file paths.

 required
decoy_prefix str

The prefix used to indicate a decoy protein in the description lines of the FASTA file. Default is "DECOY_".

 'DECOY_'

Returns:

Type Description
PsmContainer

A PsmContainer object containing the PSM data.

Raises:

Type Description
ValueError

If the PepXML files were generated by Percolator or PeptideProphet.
Notes

This function: 1. Reads and parses PepXML files 2. Calculates mass difference features 3. Processes matched ions and complementary ions 4. Creates charge columns 5. Log-transforms p-values 6. Returns a PsmContainer with the processed data

Source code in optimhc/parser/pepxml.py 
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
def read_pepxml(pepxml_files, decoy_prefix="DECOY_"):
    """
    Read PSMs from a list of PepXML files.

    Parameters
    ----------
    pepxml_files : Union[str, List[str]]
        The file path to the PepXML file or a list of file paths.
    decoy_prefix : str, optional
        The prefix used to indicate a decoy protein in the description lines
        of the FASTA file. Default is "DECOY_".

    Returns
    -------
    PsmContainer
        A PsmContainer object containing the PSM data.

    Raises
    ------
    ValueError
        If the PepXML files were generated by Percolator or PeptideProphet.

    Notes
    -----
    This function:
    1. Reads and parses PepXML files
    2. Calculates mass difference features
    3. Processes matched ions and complementary ions
    4. Creates charge columns
    5. Log-transforms p-values
    6. Returns a PsmContainer with the processed data
    """
    proton = 1.00727646677
    if isinstance(pepxml_files, str):
        pepxml_files = [pepxml_files]
    psms = pd.concat([_parse_pepxml(f, decoy_prefix) for f in pepxml_files])

    # Check that these PSMs are not from Percolator or PeptideProphet:
    illegal_cols = {
        "Percolator q-Value",
        "Percolator PEP",
        "Percolator SVMScore",
    }

    if illegal_cols.intersection(set(psms.columns)):
        raise ValueError(
            "The PepXML files appear to have generated by Percolator or "
            "PeptideProphet; hence, they should not be analyzed with mokapot."
        )

    # For open modification searches:
    psms["mass_diff"] = psms["exp_mass"] - psms["calc_mass"]
    # Calculate massdiff features
    exp_mz = psms["exp_mass"] / psms["charge"] + proton
    calc_mz = psms["calc_mass"] / psms["charge"] + proton
    psms["abs_mz_diff"] = (exp_mz - calc_mz).abs()

    # Calculate matched ions and complementary ions
    if "num_matched_ions" in psms.columns and "tot_num_ions" in psms.columns:
        if (psms["tot_num_ions"] != 0).all():
            psms["matched_ions_ratio"] = psms["num_matched_ions"] / psms["tot_num_ions"]

    # Log number of candidates:
    if "num_matched_peptides" in psms.columns:
        psms["num_matched_peptides"] = np.log10(psms["num_matched_peptides"])

    # Create charge columns:
    psms = pd.concat([psms, pd.get_dummies(psms["charge"], prefix="charge")], axis=1)

    # psms = psms.drop("charge", axis=1)
    # -log10 p-values
    nonfeat_cols = [
        "ms_data_file",
        "scan",
        "spectrum",
        "label",
        "calc_mass",  # Retain calc_mass for FlashLFQ compatibility
        "peptide",
        "proteins",
        "charge",
        "retention_time",
    ]

    feat_cols = [c for c in psms.columns if c not in nonfeat_cols]
    psms = psms.apply(_log_features, features=feat_cols)
    rescoring_features = {"Original": feat_cols}

    return PsmContainer(
        psms=psms,
        label_column="label",
        scan_column="scan",
        spectrum_column="spectrum",
        ms_data_file_column="ms_data_file",
        peptide_column="peptide",
        protein_column="proteins",
        charge_column="charge",
        rescoring_features=rescoring_features,
        hit_rank_column="rank",
        retention_time_column="retention_time",
        calculated_mass_column="calc_mass",
    )

_parse_pepxml(pepxml_file, decoy_prefix)

Parse the PSMs of a PepXML into a DataFrame.

Parameters:

Name Type Description Default
pepxml_file str

The PepXML file to parse.

 required
decoy_prefix str

The prefix used to indicate a decoy protein in the description lines of the FASTA file.

 required

Returns:

Type Description
DataFrame

A DataFrame containing the information about each PSM.

Raises:

Type Description
ValueError

If the file is not a PepXML file or is malformed.
Source code in optimhc/parser/pepxml.py 
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
def _parse_pepxml(pepxml_file, decoy_prefix):
    """
    Parse the PSMs of a PepXML into a DataFrame.

    Parameters
    ----------
    pepxml_file : str
        The PepXML file to parse.
    decoy_prefix : str
        The prefix used to indicate a decoy protein in the description lines
        of the FASTA file.

    Returns
    -------
    pandas.DataFrame
        A DataFrame containing the information about each PSM.

    Raises
    ------
    ValueError
        If the file is not a PepXML file or is malformed.
    """
    logger.info("Reading %s...", pepxml_file)
    parser = etree.iterparse(str(pepxml_file), tag="{*}msms_run_summary")
    parse_fun = partial(_parse_msms_run, decoy_prefix=decoy_prefix)
    spectra = map(parse_fun, parser)
    try:
        psms = itertools.chain.from_iterable(spectra)
        df = pd.DataFrame.from_records(itertools.chain.from_iterable(psms))
        df["ms_data_file"] = df["ms_data_file"].astype("category")
    except etree.XMLSyntaxError:
        raise ValueError(f"{pepxml_file} is not a PepXML file or is malformed.")
    return df

_parse_msms_run(msms_run, decoy_prefix)

Parse a single MS/MS run.

Parameters:

Name Type Description Default
msms_run tuple of anything, lxml.etree.Element

The second element of the tuple should be the XML element for a single msms_run. The first is not used, but is necessary for compatibility with using :code:map().

 required
decoy_prefix str

The prefix used to indicate a decoy protein in the description lines of the FASTA file.

 required

Yields:

Type Description
dict

A dictionary describing all of the PSMs in a run.
Source code in optimhc/parser/pepxml.py 
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
def _parse_msms_run(msms_run, decoy_prefix):
    """
    Parse a single MS/MS run.

    Parameters
    ----------
    msms_run : tuple of anything, lxml.etree.Element
        The second element of the tuple should be the XML element for a single
        msms_run. The first is not used, but is necessary for compatibility
        with using :code:`map()`.
    decoy_prefix : str
        The prefix used to indicate a decoy protein in the description lines
        of the FASTA file.

    Yields
    ------
    dict
        A dictionary describing all of the PSMs in a run.
    """
    msms_run = msms_run[1]
    ms_data_file = msms_run.get("base_name")
    run_ext = msms_run.get("raw_data")
    if not ms_data_file.endswith(run_ext):
        ms_data_file += run_ext

    run_info = {"ms_data_file": ms_data_file}
    for spectrum in msms_run.iter("{*}spectrum_query"):
        yield _parse_spectrum(spectrum, run_info, decoy_prefix)

_parse_spectrum(spectrum, run_info, decoy_prefix)

Parse the PSMs for a single mass spectrum.

Parameters:

Name Type Description Default
spectrum Element

The XML element for a single spectrum.

 required
run_info dict

The parsed run data.

 required
decoy_prefix str

The prefix used to indicate a decoy protein in the description lines of the FASTA file.

 required

Yields:

Type Description
dict

A dictionary describing all of the PSMs for a spectrum.
Source code in optimhc/parser/pepxml.py 
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
def _parse_spectrum(spectrum, run_info, decoy_prefix):
    """
    Parse the PSMs for a single mass spectrum.

    Parameters
    ----------
    spectrum : lxml.etree.Element
        The XML element for a single spectrum.
    run_info : dict
        The parsed run data.
    decoy_prefix : str
        The prefix used to indicate a decoy protein in the description lines
        of the FASTA file.

    Yields
    ------
    dict
        A dictionary describing all of the PSMs for a spectrum.
    """
    spec_info = run_info.copy()
    spec_info["spectrum"] = str(spectrum.get("spectrum"))
    spec_info["scan"] = int(spectrum.get("end_scan"))
    spec_info["charge"] = int(spectrum.get("assumed_charge"))
    spec_info["retention_time"] = float(spectrum.get("retention_time_sec"))
    spec_info["exp_mass"] = float(spectrum.get("precursor_neutral_mass"))
    for psms in spectrum.iter("{*}search_result"):
        for psm in psms.iter("{*}search_hit"):
            yield _parse_psm(psm, spec_info, decoy_prefix=decoy_prefix)

_parse_psm(psm_info, spec_info, decoy_prefix)

Parse a single PSM.

Parameters:

Name Type Description Default
psm_info Element

The XML element containing information about the PSM.

 required
spec_info dict

The parsed spectrum data.

 required
decoy_prefix str

The prefix used to indicate a decoy protein in the description lines of the FASTA file.

 required

Returns:

Type Description
dict

A dictionary containing parsed data about the PSM.
Source code in optimhc/parser/pepxml.py 
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
def _parse_psm(psm_info, spec_info, decoy_prefix):
    """
    Parse a single PSM.

    Parameters
    ----------
    psm_info : lxml.etree.Element
        The XML element containing information about the PSM.
    spec_info : dict
        The parsed spectrum data.
    decoy_prefix : str
        The prefix used to indicate a decoy protein in the description lines
        of the FASTA file.

    Returns
    -------
    dict
        A dictionary containing parsed data about the PSM.
    """
    psm = spec_info.copy()
    psm["calc_mass"] = float(psm_info.get("calc_neutral_pep_mass"))
    psm["peptide"] = psm_info.get("peptide")
    psm["proteins"] = [psm_info.get("protein").split(" ")[0]]
    psm["label"] = not psm["proteins"][0].startswith(decoy_prefix)
    psm["rank"] = int(psm_info.get("hit_rank"))

    # Begin features:
    try:
        psm["missed_cleavages"] = int(psm_info.get("num_missed_cleavages"))
    except TypeError:
        pass

    try:
        psm["ntt"] = int(psm_info.get("num_tol_term"))
    except TypeError:
        pass

    try:
        psm["num_matched_peptides"] = int(psm_info.get("num_matched_peptides"))
    except TypeError:
        pass

    try:
        psm["num_matched_ions"] = int(psm_info.get("num_matched_ions"))
    except TypeError:
        pass

    try:
        psm["tot_num_ions"] = int(psm_info.get("tot_num_ions"))
    except TypeError:
        pass

    queries = [
        "{*}modification_info",
        "{*}search_score",
        "{*}alternative_protein",
    ]

    # TODO: Directly read modification_info
    # TODO: The current code can not parse the n-terminal modification
    for element in psm_info.iter(*queries):
        if "modification_info" in element.tag:
            offset = 0
            mod_pep = psm["peptide"]
            for mod in element.iter("{*}mod_aminoacid_mass"):
                idx = offset + int(mod.get("position"))
                mass = mod.get("mass")
                mod_pep = mod_pep[:idx] + "[" + mass + "]" + mod_pep[idx:]
                offset += 2 + len(mass)

            psm["peptide"] = mod_pep

        elif "alternative_protein" in element.tag:
            psm["proteins"].append(element.get("protein").split(" ")[0])
            if not psm["label"]:
                psm["label"] = not psm["proteins"][-1].startswith(decoy_prefix)

        else:
            psm[element.get("name")] = element.get("value")

    psm["proteins"] = "\t".join(psm["proteins"])
    return psm

_log_features(col, features)

Log-transform columns that are p-values or E-values.

Parameters:

Name Type Description Default
col Series

A column of the dataset.

 required
features list of str

The features of the dataset. Only feature columns will be considered for transformation.

 required

Returns:

Type Description
Series

The log-transformed values of the column if the feature was determined to be a p-value.
Notes

This function: 1. Detects columns written in scientific notation and log them 2. Uses a simple heuristic to find p-value / E-value features 3. Only transforms if values span >4 orders of magnitude 4. Preserves precision for scientific notation values

Source code in optimhc/parser/pepxml.py 
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
def _log_features(col, features):
    """
    Log-transform columns that are p-values or E-values.

    Parameters
    ----------
    col : pandas.Series
        A column of the dataset.
    features : list of str
        The features of the dataset. Only feature columns will be considered
        for transformation.

    Returns
    -------
    pandas.Series
        The log-transformed values of the column if the feature was determined
        to be a p-value.

    Notes
    -----
    This function:
    1. Detects columns written in scientific notation and log them
    2. Uses a simple heuristic to find p-value / E-value features
    3. Only transforms if values span >4 orders of magnitude
    4. Preserves precision for scientific notation values
    """
    if col.name not in features:
        return col
    elif col.dtype == "bool":
        return col.astype(float)

    col = col.astype(str).str.lower()

    # Detect columns written in scientific notation and log them:
    # This is specifically needed to preserve precision.
    if col.str.contains("e").any() and (col.astype(float) > 0).all():
        split = col.str.split("e", expand=True)
        root = split.loc[:, 0]
        root = root.astype(float)
        power = split.loc[:, 1]
        power[pd.isna(power)] = "0"
        power = power.astype(int)

        zero_idx = root == 0
        root[zero_idx] = 1
        power[zero_idx] = power[~zero_idx].min()
        diff = power.max() - power.min()
        if abs(diff) >= 4:
            logger.info("  - log-transformed the '%s' feature.", col.name)
            return np.log10(root) + power
        else:
            return col.astype(float)

    col = col.astype(float)

    # A simple heuristic to find p-value / E-value features:
    # Non-negative:
    if col.min() >= 0:
        # Make sure this isn't a binary column:
        if not np.array_equal(col.values, col.values.astype(bool)):
            # Only log if values span >4 orders of magnitude,
            # excluding values that are exactly zero:
            zero_idx = col == 0
            col_min = col[~zero_idx].min()
            if col.max() / col_min >= 10000:
                col[~zero_idx] = np.log10(col[~zero_idx])
                col[zero_idx] = col[~zero_idx].min() - 1
                logger.info("  - log-transformed the '%s' feature.", col.name)

    return col

PIN Parser

pin

logger = logging.getLogger(__name__) module-attribute

PsmContainer(psms, label_column, scan_column, spectrum_column, ms_data_file_column, peptide_column, protein_column, rescoring_features, hit_rank_column=None, charge_column=None, retention_time_column=None, calculated_mass_column=None, metadata_column=None)

A container for managing peptide-spectrum matches (PSMs) in immunopeptidomics rescoring pipelines.

Parameters:

Name Type Description Default
psms DataFrame

DataFrame containing the PSM data.

 required
label_column str

Column containing the label (True for target, False for decoy).

 required
scan_column str

Column containing the scan number.

 required
spectrum_column str

Column containing the spectrum identifier.

 required
ms_data_file_column str

Column containing the MS data file that the PSM originated from.

 required
peptide_column str

Column containing the peptide sequence.

 required
protein_column str

Column containing the protein accessions.

 required
rescoring_features dict of str to list of str

Dictionary of feature columns for rescoring.

 required
hit_rank_column str

Column containing the hit rank.

 None
charge_column str

Column containing the charge state.

 None
retention_time_column str

Column containing the retention time.

 None
calculated_mass_column str

Column containing the calculated mass.

 None
metadata_column str

Column containing metadata.

 None

Attributes:

Name Type Description
psms DataFrame

Copy of the DataFrame containing the PSM data.
target_psms DataFrame

DataFrame containing only target PSMs (label = True).
decoy_psms DataFrame

DataFrame containing only decoy PSMs (label = False).
peptides list of str

List containing all peptides from the PSM data.
columns list of str

List of column names in the PSM DataFrame.
rescoring_features dict of str to list of str

Dictionary of rescoring feature columns in the PSM DataFrame.
Source code in optimhc/psm_container.py 
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
def __init__(
    self,
    psms: pd.DataFrame,
    label_column: str,
    scan_column: str,
    spectrum_column: str,
    ms_data_file_column: str,
    peptide_column: str,
    protein_column: str,
    rescoring_features: Dict[str, List[str]],
    hit_rank_column: Optional[str] = None,
    charge_column: Optional[str] = None,
    retention_time_column: Optional[str] = None,
    calculated_mass_column: Optional[str] = None,
    metadata_column: Optional[str] = None,
):
    self._psms = psms.copy()
    self._psms.reset_index(drop=True, inplace=True)
    self.label_column = label_column
    self.scan_column = scan_column
    self.spectrum_column = spectrum_column
    self.ms_data_file_column = ms_data_file_column
    self.peptide_column = peptide_column
    self.protein_column = protein_column
    self.hit_rank_column = hit_rank_column
    self.retention_time_column = retention_time_column
    self.metadata_column = metadata_column
    self.rescoring_features = rescoring_features
    self.charge_column = charge_column
    self.calculated_mass_column = calculated_mass_column
    # rescore result column
    self.rescore_result_column = None

    # check if the columns are in the dataframe
    def check_column(col):
        if col and col not in psms.columns:
            raise ValueError(f"Column '{col}' not found in PSM data.")

    check_column(label_column)
    check_column(scan_column)
    check_column(spectrum_column)
    check_column(ms_data_file_column)
    check_column(peptide_column)
    check_column(protein_column)
    check_column(hit_rank_column)
    check_column(retention_time_column)
    check_column(charge_column)
    check_column(calculated_mass_column)

    # ensure the label column is boolean
    if psms[label_column].dtype != "bool":
        raise ValueError(f"Column '{label_column}' must be boolean.")

    if psms[label_column].nunique() == 1 and psms[label_column].iloc[0]:
        raise ValueError("All PSMs are labeled as target. No decoy PSMs found.")
    elif psms[label_column].nunique() == 1 and not psms[label_column].iloc[0]:
        raise ValueError("All PSMs are labeled as decoy. No target PSMs found.")

    def check_metadata_column(col):
        # check the type is Dict[str, Dict[str, str]]
        if col and col not in psms.columns:
            raise ValueError(f"Column '{col}' not found in PSM data.")
        if not all(isinstance(x, dict) for x in self._psms[col]):
            raise ValueError(f"Column '{col}' must contain dictionaries.")

    if metadata_column:
        check_metadata_column(metadata_column)

    def check_rescoring_features(features: Dict[str, List[str]]):
        for key, cols in features.items():
            for col in cols:
                if col not in psms.columns:
                    raise ValueError(
                        f"Column '{col}' not found in PSM data for feature '{key}'."
                    )

    check_rescoring_features(rescoring_features)

    # check if the number of decoy psms is not 0
    if len(self.decoy_psms) == 0:
        logger.error("No decoy PSMs found. Please check the decoy prefix.")
        raise ValueError("No decoy PSMs found.")

    logger.info("PsmContainer initialized with %d PSM entries.", len(self._psms))
    if self.ms_data_file_column:
        logger.info(
            "PSMs originated from %d MS data file(s).",
            len(self._psms[ms_data_file_column].unique()),
        )
    logger.info("target psms: %d", len(self.target_psms))
    logger.info("decoy psms: %d", len(self.decoy_psms))
    logger.info("unique peptides: %d", len(np.unique(self.peptides)))
    logger.info("rescoring features: %s", rescoring_features)

psms property

Get a copy of the PSM DataFrame to prevent external modification.

Returns:

Type Description
DataFrame

A copy of the PSM DataFrame.

target_psms property

Get a DataFrame containing only target PSMs.

Returns:

Type Description
DataFrame

DataFrame with only target PSMs (label = True).

decoy_psms property

Get a DataFrame containing only decoy PSMs.

Returns:

Type Description
DataFrame

DataFrame with only decoy PSMs (label = False).

columns property

Get the column names of the PSM DataFrame.

Returns:

Type Description
list of str

List of column names.

feature_columns property

Get a list of all feature columns in the PSM DataFrame.

Returns:

Type Description
list of str

List of feature column names.

feature_sources property

Get a list of all feature sources in the PSM DataFrame.

Returns:

Type Description
list of str

List of feature source names.

peptides property

Get the peptide sequences from the PSM data.

Returns:

Type Description
list of str

List of peptide sequences.

ms_data_files property

Get the MS data files from the PSM data.

Returns:

Type Description
list of str

List of MS data file names.

scan_ids property

Get the scan numbers from the PSM data.

Returns:

Type Description
list of int

List of scan numbers.

charges property

Get the charge states from the PSM data.

Returns:

Type Description
list of int

List of charge states.

metadata property

Get the metadata from the PSM data.

Returns:

Type Description
Series

Series containing metadata for each PSM.

spectrum_ids property

Get the spectrum identifiers from the PSM data.

Returns:

Type Description
list of str

List of spectrum identifiers.

identifier_columns property

Get the columns that uniquely identify each PSM.

Returns:

Type Description
list of str

List of identifier column names.

__len__()

Get the number of PSMs in the container.

Returns:

Type Description
int

Number of PSMs.
Source code in optimhc/psm_container.py 
168
169
170
171
172
173
174
175
176
177
def __len__(self) -> int:
    """
    Get the number of PSMs in the container.

    Returns
    -------
    int
        Number of PSMs.
    """
    return len(self._psms)

copy()

Return a deep copy of the PsmContainer object.

Returns:

Type Description
PsmContainer

A deep copy of the current PsmContainer.
Source code in optimhc/psm_container.py 
328
329
330
331
332
333
334
335
336
337
338
339
def copy(self) -> "PsmContainer":
    """
    Return a deep copy of the PsmContainer object.

    Returns
    -------
    PsmContainer
        A deep copy of the current PsmContainer.
    """
    import copy

    return copy.deepcopy(self)

__repr__()

Return a string representation of the PsmContainer.

Returns:

Type Description
str

String summary of the PsmContainer.
Source code in optimhc/psm_container.py 
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
def __repr__(self) -> str:
    """
    Return a string representation of the PsmContainer.

    Returns
    -------
    str
        String summary of the PsmContainer.
    """
    return (
        f"PsmContainer with {len(self)} PSMs\n"
        f"\t - Target PSMs: {len(self.target_psms)}\n"
        f"\t - Decoy PSMs: {len(self.decoy_psms)}\n"
        f"\t - Unique Peptides: {len(np.unique(self.peptides))}\n"
        f"\t - Unique Spectra: {len(self._psms[self.spectrum_column].unique())}\n"
        f"\t - Rescoring Features: {self.rescoring_features}\n"
    )

drop_features(features)

Drop specified features from the PSM DataFrame.

Parameters:

Name Type Description Default
features list of str

List of feature column names to drop.

 required

Raises:

Type Description
ValueError

If any of the features do not exist in the DataFrame.
Source code in optimhc/psm_container.py 
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
def drop_features(self, features: List[str]) -> None:
    """
    Drop specified features from the PSM DataFrame.

    Parameters
    ----------
    features : list of str
        List of feature column names to drop.

    Raises
    ------
    ValueError
        If any of the features do not exist in the DataFrame.
    """
    missing_features = [f for f in features if f not in self._psms.columns]
    if missing_features:
        raise ValueError(f"Features not found in PSM data: {missing_features}")

    self._psms.drop(columns=features, inplace=True)
    # Create a list of sources to update
    sources_to_update = []
    for source, cols in self.rescoring_features.items():
        self.rescoring_features[source] = [col for col in cols if col not in features]
        if not self.rescoring_features[source]:
            sources_to_update.append(source)

    logger.info(
        f"Sources to be removed: {sources_to_update}. Because all the features are removed."
    )
    # Remove sources with no features left
    for source in sources_to_update:
        del self.rescoring_features[source]

drop_source(source)

Drop all features associated with a specific source from the PSM DataFrame.

Parameters:

Name Type Description Default
source str

Name of the source to drop.

 required

Raises:

Type Description
ValueError

If the source does not exist in the rescoring features.
Source code in optimhc/psm_container.py 
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
def drop_source(self, source: str) -> None:
    """
    Drop all features associated with a specific source from the PSM DataFrame.

    Parameters
    ----------
    source : str
        Name of the source to drop.

    Raises
    ------
    ValueError
        If the source does not exist in the rescoring features.
    """
    if source not in self.rescoring_features:
        raise ValueError(f"Source '{source}' not found in rescoring features.")
    self.drop_features(self.rescoring_features[source])

add_metadata(metadata_df, psms_key, metadata_key, source)

Merge new metadata into the PSM DataFrame based on specified columns. Metadata from the specified source is stored as a nested dictionary inside the metadata column.

Parameters:

Name Type Description Default
metadata_df DataFrame

DataFrame containing new metadata to add.

 required
psms_key str or list of str

Column name(s) in the PSM data to merge on.

 required
metadata_key str or list of str

Column name(s) in the metadata data to merge on.

 required
source str

Name of the source of the new metadata.

 required
Source code in optimhc/psm_container.py 
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
def add_metadata(
    self,
    metadata_df: pd.DataFrame,
    psms_key: Union[str, List[str]],
    metadata_key: Union[str, List[str]],
    source,
) -> None:
    """
    Merge new metadata into the PSM DataFrame based on specified columns.
    Metadata from the specified source is stored as a nested dictionary inside the metadata column.

    Parameters
    ----------
    metadata_df : pd.DataFrame
        DataFrame containing new metadata to add.
    psms_key : str or list of str
        Column name(s) in the PSM data to merge on.
    metadata_key : str or list of str
        Column name(s) in the metadata data to merge on.
    source : str
        Name of the source of the new metadata.
    """
    if self.metadata_column is None:
        logger.info("No existing metadata column. Creating new metadata column.")
        self.metadata_column = "metadata"
        self._psms["metadata"] = [{} for _ in range(len(self._psms))]

    metadata_cols = [col for col in metadata_df.columns if col not in metadata_key]
    merged_df = self.psms.merge(
        metadata_df, left_on=psms_key, right_on=metadata_key, how="left"
    )
    if source in self._psms["metadata"]:
        logger.warning(f"{source} already exists in metadata. Overwriting.")
    for col in metadata_cols:
        merged_df["metadata"] = merged_df.apply(
            lambda row: {
                **row["metadata"],
                source: (
                    {col: row[col]}
                    if source not in row["metadata"]
                    else {**row["metadata"][source], col: row[col]}
                ),
            },
            axis=1,
        )

    self._psms["metadata"] = merged_df["metadata"]

get_top_hits(n=1)

Get the top n hits based on the hit rank column. If the hit rank column is not specified, returns the original PSMs.

Parameters:

Name Type Description Default
n int

The number of top hits to return. Default is 1.

 1

Returns:

Type Description
PsmContainer

A new PsmContainer object containing the top n hits.
Source code in optimhc/psm_container.py 
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
def get_top_hits(self, n: int = 1):
    """
    Get the top n hits based on the hit rank column.
    If the hit rank column is not specified, returns the original PSMs.

    Parameters
    ----------
    n : int, optional
        The number of top hits to return. Default is 1.

    Returns
    -------
    PsmContainer
        A new PsmContainer object containing the top n hits.
    """
    if self.hit_rank_column is None:
        logger.warning("Rank column not specified. Return the original PSMs.")
        return self.copy()

    psms = self.copy()
    psms._psms = psms._psms[psms._psms[self.hit_rank_column] <= n]
    return psms

add_features(features_df, psms_key, feature_key, source, suffix=None)

Merge new features into the PSM DataFrame based on specified columns.

This method performs a left join between the PSM data and feature data, ensuring that all PSMs are preserved while adding new features. It handles column name conflicts through optional suffixing and maintains feature source tracking.

Parameters:

Name Type Description Default
features_df DataFrame

DataFrame containing new features to add.

 required
psms_key str or list of str

Column name(s) in the PSM data to merge on.

 required
feature_key str or list of str

Column name(s) in the features data to merge on.

 required
source str

Name of the source of the new features (e.g., 'deeplc', 'netmhc').

 required
suffix str

Suffix to add to the new columns if there's a name conflict. Required when new feature columns have the same names as existing columns. For example, if adding features from different sources (e.g., 'score' from DeepLC and NetMHC), use suffixes like '_deeplc' or '_netmhc' to distinguish them.

 None

Returns:

Type Description
None

Raises:

Type Description
ValueError

If duplicate columns exist without suffix. If merging features changes the number of PSMs.
Notes

The method follows these steps: 1. Validates input and prepares merge keys 2. Checks for column name conflicts 3. Manages feature source: if the source already exists, it will be overwritten 4. Performs left join merge 5. Verifies data integrity

Suffix Usage

The suffix parameter is used to handle column name conflicts: - When adding features from different sources that might have the same column names - When you want to keep both the original and new features with the same name - When you need to track the source of features in the column names

If suffix is not provided and there are duplicate column names: - The method will raise a ValueError - You must either provide a suffix or rename the columns before adding

Examples:

>>> container = PsmContainer(...)
>>> # Adding features without suffix (no conflicts)
>>> features_df1 = pd.DataFrame({
...     'scan': [1, 2, 3],
...     'feature1': [0.1, 0.2, 0.3],
...     'feature2': [0.4, 0.5, 0.6]
... })
>>> container.add_features(
...     features_df1,
...     psms_key='scan',
...     feature_key='scan',
...     source='source1'
... )
>>> # Adding features with suffix (handling conflicts)
>>> features_df2 = pd.DataFrame({
...     'scan': [1, 2, 3],
...     'score': [0.8, 0.9, 0.7],  # This would conflict with existing 'score'
...     'feature3': [0.7, 0.8, 0.9]
... })
>>> container.add_features(
...     features_df2,
...     psms_key='scan',
...     feature_key='scan',
...     source='source2',
...     suffix='_new'  # 'score' becomes 'score_new'
... )
Source code in optimhc/psm_container.py 
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
def add_features(
    self,
    features_df: pd.DataFrame,
    psms_key: Union[str, List[str]],
    feature_key: Union[str, List[str]],
    source: str,
    suffix: Optional[str] = None,
) -> None:
    """Merge new features into the PSM DataFrame based on specified columns.

    This method performs a left join between the PSM data and feature data,
    ensuring that all PSMs are preserved while adding new features. It handles
    column name conflicts through optional suffixing and maintains feature source
    tracking.

    Parameters
    ----------
    features_df : pd.DataFrame
        DataFrame containing new features to add.
    psms_key : str or list of str
        Column name(s) in the PSM data to merge on.
    feature_key : str or list of str
        Column name(s) in the features data to merge on.
    source : str
        Name of the source of the new features (e.g., 'deeplc', 'netmhc').
    suffix : str, optional
        Suffix to add to the new columns if there's a name conflict.
        Required when new feature columns have the same names as existing columns.
        For example, if adding features from different sources (e.g., 'score' from
        DeepLC and NetMHC), use suffixes like '_deeplc' or '_netmhc' to distinguish them.

    Returns
    -------
    None

    Raises
    ------
    ValueError
        If duplicate columns exist without suffix.
        If merging features changes the number of PSMs.

    Notes
    -----
    The method follows these steps:
    1. Validates input and prepares merge keys
    2. Checks for column name conflicts
    3. Manages feature source: if the source already exists, it will be overwritten
    4. Performs left join merge
    5. Verifies data integrity

    Suffix Usage
    -----------
    The suffix parameter is used to handle column name conflicts:
    - When adding features from different sources that might have the same column names
    - When you want to keep both the original and new features with the same name
    - When you need to track the source of features in the column names

    If suffix is not provided and there are duplicate column names:
    - The method will raise a ValueError
    - You must either provide a suffix or rename the columns before adding

    Examples
    --------
    >>> container = PsmContainer(...)
    >>> # Adding features without suffix (no conflicts)
    >>> features_df1 = pd.DataFrame({
    ...     'scan': [1, 2, 3],
    ...     'feature1': [0.1, 0.2, 0.3],
    ...     'feature2': [0.4, 0.5, 0.6]
    ... })
    >>> container.add_features(
    ...     features_df1,
    ...     psms_key='scan',
    ...     feature_key='scan',
    ...     source='source1'
    ... )
    >>> # Adding features with suffix (handling conflicts)
    >>> features_df2 = pd.DataFrame({
    ...     'scan': [1, 2, 3],
    ...     'score': [0.8, 0.9, 0.7],  # This would conflict with existing 'score'
    ...     'feature3': [0.7, 0.8, 0.9]
    ... })
    >>> container.add_features(
    ...     features_df2,
    ...     psms_key='scan',
    ...     feature_key='scan',
    ...     source='source2',
    ...     suffix='_new'  # 'score' becomes 'score_new'
    ... )
    """
    if isinstance(psms_key, str):
        psms_key = [psms_key]

    if isinstance(feature_key, str):
        feature_key = [feature_key]

    new_feature_cols = [col for col in features_df.columns if col not in feature_key]

    for cols in new_feature_cols:
        if cols in self._psms.columns:
            logger.warning(f"Column '{cols}' already exists in PSM data.")
            if suffix is None:
                logger.warning("No suffix provided. Using default suffix ")
                raise ValueError("Duplicate columns exist. No suffix provided.")
            else:
                logger.warning(f"Suffix '{suffix}' provided. Using suffix '{suffix}'.")
    logger.info(f"Adding {len(new_feature_cols)} new features from {source}.")

    if not new_feature_cols:
        logger.warning("No new features to add. Check the feature key and PSMs key.")
        logger.warning(f"Feature key: {feature_key}; PSMs key: {psms_key}")

    if source in self.rescoring_features:
        logger.warning(f"{source} already exists in rescoring features. Overwriting.")
        self.drop_source(source)

    # TODO: reluctant logic
    if suffix is None:
        suffixes = ("", "")
    else:
        suffixes = ("", suffix)

    self.rescoring_features[source] = [col + suffixes[1] for col in new_feature_cols]
    features_df = features_df.rename(
        columns={col: col + suffixes[1] for col in new_feature_cols}
    )
    original_len = len(self._psms)
    # avoid merge the right key to the psms
    self._psms = self._psms.merge(
        features_df, left_on=psms_key, right_on=feature_key, how="left"
    )

    if feature_key != psms_key:
        cols_to_drop = [
            col for col in feature_key if col not in psms_key and col in self._psms.columns
        ]
        if cols_to_drop:
            logger.debug(f"Dropping columns from feature_key not in psms_key: {cols_to_drop}")
            self._psms.drop(columns=cols_to_drop, inplace=True)

    if len(self._psms) != original_len:
        raise ValueError(
            "Merging features resulted in a change in the number of PSMs. Check for duplicate keys."
        )

add_features_by_index(features_df, source, suffix=None)

Merge new features into the PSM DataFrame based on the DataFrame index.

Parameters:

Name Type Description Default
features_df DataFrame

DataFrame containing new features to add.

 required
source str

Name of the source of the new features.

 required
suffix str

Suffix to add to the new columns if there's a name conflict.

 None
Source code in optimhc/psm_container.py 
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
def add_features_by_index(
    self, features_df: pd.DataFrame, source: str, suffix: Optional[str] = None
) -> None:
    """
    Merge new features into the PSM DataFrame based on the DataFrame index.

    Parameters
    ----------
    features_df : pd.DataFrame
        DataFrame containing new features to add.
    source : str
        Name of the source of the new features.
    suffix : str, optional
        Suffix to add to the new columns if there's a name conflict.
    """
    new_feature_cols = [col for col in features_df.columns]
    for col in new_feature_cols:
        if col in self._psms.columns:
            logger.warning(f"Column '{col}' already exists in PSM data.")
            if suffix is None:
                logger.warning("No suffix provided. Using default suffix.")
                raise ValueError("Duplicate columns exist. No suffix provided.")
            else:
                logger.warning(f"Suffix '{suffix}' provided. Using suffix '{suffix}'.")

    logger.info(f"Adding {len(new_feature_cols)} new features from {source} by index.")

    if not new_feature_cols:
        logger.warning("No new features to add.")
        raise ValueError("No new features to add.")

    if source in self.rescoring_features:
        logger.warning(f"{source} already exists in rescoring features. Overwriting.")
        self.drop_source(source)

    if suffix is None:
        suffixes = ("", "")
    else:
        suffixes = ("", suffix)

    self.rescoring_features[source] = [col + suffixes[1] for col in new_feature_cols]
    features_df.rename(
        columns={col: col + suffixes[1] for col in new_feature_cols}, inplace=True
    )
    original_len = len(self._psms)
    self._psms = self._psms.merge(
        features_df,
        left_index=True,
        right_index=True,
        how="left",  # Perform a left join to preserve all original PSM data
    )

    # Ensure that the merge did not change the number of rows in the PSM DataFrame
    if len(self._psms) != original_len:
        raise ValueError(
            "Merging features resulted in a change in the number of PSMs. Check for duplicate indices."
        )

add_results(results_df, psms_key, result_key)

Add results of rescore engine to the PSM DataFrame based on specified columns.

Parameters:

Name Type Description Default
results_df DataFrame

DataFrame containing new results to add.

 required
psms_key str or list of str

Column name(s) in the PSM data to merge on.

 required
result_key str or list of str

Column name(s) in the results data to merge on.

 required
Source code in optimhc/psm_container.py 
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
def add_results(
    self,
    results_df: pd.DataFrame,
    psms_key: Union[str, List[str]],
    result_key: Union[str, List[str]],
) -> None:
    """
    Add results of rescore engine to the PSM DataFrame based on specified columns.

    Parameters
    ----------
    results_df : pd.DataFrame
        DataFrame containing new results to add.
    psms_key : str or list of str
        Column name(s) in the PSM data to merge on.
    result_key : str or list of str
        Column name(s) in the results data to merge on.
    """
    if self.rescore_result_column is not None:
        logger.warning("Rescore result column already exists. Overwriting.")

    if set(self._psms.columns) & set(results_df.columns):
        raise ValueError(
            "Duplicate columns exist. Please rename the columns in the results data."
        )

    self.rescore_result_column = result_key
    self._psms = self._psms.merge(
        results_df,
        left_on=psms_key,
        right_on=result_key,
        how="left",
        validate="one_to_one",
    )
    self._psms.drop(columns=result_key, inplace=True)
    logger.info("Added rescore results to PSM data.")

write_pin(output_file, style='default', source=None)

Write the PSM data to a Percolator PIN file, supporting both generic Percolator and MSBooster-compatible formats. The style parameter is actually used to output a unified pin format file to benchmark the performance of different rescoring methods.

Parameters:

Name Type Description Default
output_file str

Path to the output PIN file.

 required
style str

If set to 'msbooster', outputs only the columns required by MSBooster (SpecId, Label, ScanNr, retentiontime, rank, hyperscore or log10_evalue, Peptide, Proteins). If set to 'default', outputs all features specified in rescoring_features, plus required Percolator columns.

 'default'
source list of str

List of feature sources to include. If None, includes all sources.

 None

Returns:

Type Description
DataFrame

The DataFrame written to the PIN file.
Notes
  • The first three columns are always: SpecID, Label, ScanNr.
  • For 'msbooster' style, the columns are: SpecId, Label, ScanNr, retentiontime, rank, hyperscore or log10_evalue, Peptide, Proteins.
  • If hit_rank_column is not specified, rank is set to 1 for all rows.
  • Either 'hyperscore' or 'expect' must be present in features; for 'expect', the column is written as 'log10_evalue'.
  • The 'log10_evalue' column should contain the base-10 logarithm of the e-value.
  • The 'Peptide' column is formatted with underscores (e.g., _.PEPTIDE._).
  • For standard format, all features from rescoring_features are appended between ScanNr and Peptide columns.
  • The 'Proteins' column is a semicolon-separated list if stored as a list or tuple.
  • Label column is converted to 1 (target) and -1 (decoy), as required by Percolator.

Example output (default style): SpecId Label ScanNr feature1 feature2 ... Peptide Proteins

Example output (msbooster style): SpecId Label ScanNr retentiontime rank hyperscore Peptide Proteins or SpecId Label ScanNr retentiontime rank log10_evalue Peptide Proteins

Raises:

Type Description
ValueError

If required columns are missing for the selected style.
Source code in optimhc/psm_container.py 
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
def write_pin(
    self, output_file: str, style: str = "default", source: List[str] = None
) -> None:
    """
    Write the PSM data to a Percolator PIN file, supporting both generic Percolator and MSBooster-compatible formats.
    The style parameter is actually used to output a unified pin format file to benchmark the performance of different rescoring methods.

    Parameters
    ----------
    output_file : str
        Path to the output PIN file.
    style : str, optional
        If set to 'msbooster', outputs only the columns required by MSBooster (SpecId, Label, ScanNr, retentiontime, rank, hyperscore or log10_evalue, Peptide, Proteins).
        If set to 'default', outputs all features specified in `rescoring_features`, plus required Percolator columns.
    source : list of str, optional
        List of feature sources to include. If None, includes all sources.

    Returns
    -------
    pd.DataFrame
        The DataFrame written to the PIN file.

    Notes
    -----
    - The first three columns are always: SpecID, Label, ScanNr.
    - For 'msbooster' style, the columns are: SpecId, Label, ScanNr, retentiontime, rank, hyperscore or log10_evalue, Peptide, Proteins.
    - If `hit_rank_column` is not specified, rank is set to 1 for all rows.
    - Either 'hyperscore' or 'expect' must be present in features; for 'expect', the column is written as 'log10_evalue'.
    - The 'log10_evalue' column should contain the base-10 logarithm of the e-value.
    - The 'Peptide' column is formatted with underscores (e.g., `_.PEPTIDE._`).
    - For standard format, all features from `rescoring_features` are appended between ScanNr and Peptide columns.
    - The 'Proteins' column is a semicolon-separated list if stored as a list or tuple.
    - Label column is converted to 1 (target) and -1 (decoy), as required by Percolator.

    Example output (default style):
        SpecId	Label	ScanNr	feature1	feature2	...	Peptide	Proteins

    Example output (msbooster style):
        SpecId	Label	ScanNr	retentiontime	rank	hyperscore	Peptide	Proteins
        or
        SpecId	Label	ScanNr	retentiontime	rank	log10_evalue	Peptide	Proteins

    Raises
    ------
    ValueError
        If required columns are missing for the selected style.
    """
    df = self._psms.copy()
    # Check if the label column is str
    # Case1: label column is str
    if df[self.label_column].dtype == "str":
        df["PercolatorLabel"] = df[self.label_column].map({"True": 1, "False": -1})
    # Case2: label column is bool
    elif df[self.label_column].dtype == "bool":
        df["PercolatorLabel"] = df[self.label_column].map({True: 1, False: -1})
    else:
        # try to convert to bool
        logger.warning("Label column is not str or bool. Converting to bool.")
        df["PercolatorLabel"] = df[self.label_column].astype(bool).map({True: 1, False: -1})
    logger.info("Writing PIN file to %s", output_file)
    logger.info("Using style: %s", style)

    feature_cols = []
    if source is None:
        for _, cols in self.rescoring_features.items():
            feature_cols.extend(cols)
    else:
        for s in source:
            if s not in self.rescoring_features:
                raise ValueError(f"Source '{s}' not found in rescoring features.")
            feature_cols.extend(self.rescoring_features[s])

    pin_df = pd.DataFrame()
    pin_df["SpecId"] = df[self.spectrum_column]
    pin_df["Label"] = df["PercolatorLabel"]
    pin_df["ScanNr"] = df[self.scan_column]

    if style == "msbooster":
        if self.retention_time_column:
            pin_df["retentiontime"] = df[self.retention_time_column]
        else:
            raise ValueError("Retention time column is required for msbooster style.")

        pin_df["rank"] = df[self.hit_rank_column].astype(int) if self.hit_rank_column else 1
        if "hyperscore" in self.feature_columns:
            pin_df["hyperscore"] = df["hyperscore"]
        elif "expect" in self.feature_columns:
            pin_df["log10_evalue"] = df["expect"]
        else:
            raise ValueError(
                "Either 'hyperscore' or 'expect' column is required for msbooster style."
            )

        # Add other features and jump the hyperscore or expect column
        for col in feature_cols:
            if col not in [
                "hyperscore",
                "expect",
                self.hit_rank_column,
                self.retention_time_column,
            ]:
                pin_df[col] = df[col]

        # PEPTIDE -> _.PEPTIDE._
        # Add _. at the front and ._ at the end of the peptide column
        pin_df["Peptide"] = df[self.peptide_column].apply(
            lambda x: f"_.{x}._" if isinstance(x, str) else x
        )

    elif style == "default":
        for col in feature_cols:
            pin_df[col] = df[col]
        pin_df["Peptide"] = df[self.peptide_column]
    else:
        raise ValueError(f"Unknown style: {style}. Use 'msbooster' or 'default'.")

    pin_df["Proteins"] = df[self.protein_column].apply(
        lambda x: ";".join(x) if isinstance(x, (list, tuple)) else x
    )
    pin_df = self._convert_float_to_int(pin_df)
    pin_df.to_csv(output_file, sep="\t", index=False)
    logger.info("PIN file written to %s", output_file)
    return pin_df

read_pin(pin_files, retention_time_column=None, remove_pre_nxt_aa=False)

Read PSMs from a Percolator INput (PIN) file.

Parameters:

Name Type Description Default
pin_files Union[str, List[str]]

The file path to the PIN file or a list of file paths.

 required
retention_time_column Optional[str]

The column containing the retention time. If None, no retention time will be included.

 None

Returns:

Type Description
PsmContainer

A PsmContainer object containing the PSM data.
Notes

This function: 1. Reads PIN file(s) into a DataFrame 2. Identifies required columns (case-insensitive) 3. Processes scan IDs and hit ranks (Only support FragPipe PIN) 4. Converts data types appropriately 5. Creates a PsmContainer with the processed data

Source code in optimhc/parser/pin.py 
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
def read_pin(
    pin_files: Union[str, List[str]],
    retention_time_column: Optional[str] = None,
    remove_pre_nxt_aa: bool = False,
) -> PsmContainer:
    """
    Read PSMs from a Percolator INput (PIN) file.

    Parameters
    ----------
    pin_files : Union[str, List[str]]
        The file path to the PIN file or a list of file paths.
    retention_time_column : Optional[str], optional
        The column containing the retention time. If None, no retention time
        will be included.

    Returns
    -------
    PsmContainer
        A PsmContainer object containing the PSM data.

    Notes
    -----
    This function:
    1. Reads PIN file(s) into a DataFrame
    2. Identifies required columns (case-insensitive)
    3. Processes scan IDs and hit ranks (Only support FragPipe PIN)
    4. Converts data types appropriately
    5. Creates a PsmContainer with the processed data
    """
    logger.info("Reading PIN file(s) into PsmContainer.")
    if isinstance(pin_files, str):
        pin_files = [pin_files]

    pin_df = pd.concat([_read_single_pin_as_df(pin_file) for pin_file in pin_files])
    logger.info(f"Read {len(pin_df)} PSMs from {len(pin_files)} PIN files.")
    logger.debug(pin_df.head())
    logger.debug(pin_df.columns)
    logger.debug(pin_df.iloc[0])

    def find_required_columns(col: str, columns: List[str]) -> str:
        """
        Case-insensitive search for a column in the DataFrame.
        Returns the matching column name with original casing.
        """
        col_lower = col.lower()
        column_map = {c.lower(): c for c in columns}
        if col_lower not in column_map:
            raise ValueError(f"Column '{col}' not found in PSM data (case-insensitive).")
        return column_map[col_lower]

    # non-feature columns (case-insensitive search)
    label = find_required_columns("Label", pin_df.columns)
    scan = find_required_columns("ScanNr", pin_df.columns)
    specid = find_required_columns("SpecId", pin_df.columns)
    peptide = find_required_columns("Peptide", pin_df.columns)
    protein = find_required_columns("Proteins", pin_df.columns)

    # Comet: P2PI20160713_pilling_C1RA2_BB72_P1_31_3_1
    # Fragpipe: P2PI20160713_pilling_C1RA2_BB72_P1.3104.3104.2_1

    # Try to parse rank from SpecId
    def parse_specid(specid: str) -> Tuple[str, int]:
        if "_" in specid:
            parts = specid.rsplit("_", 1)
            if len(parts) != 2:
                logger.warning(f"SpecId format unexpected: {specid}, using default rank 1")
                return 1
            try:
                hit_rank = int(parts[1])
                return hit_rank
            except ValueError:
                logger.warning(f"Could not parse rank from SpecId: {specid}, using default rank 1")
                return 1
        else:
            return 1

    hit_rank = "rank"
    if "rank" in [c.lower() for c in pin_df.columns]:
        pass
    else:
        # Parse SpecId to extract hit rank and update both columns
        pin_df["rank"] = pin_df[specid].apply(parse_specid)

    retention_time_column = (
        find_required_columns(retention_time_column, pin_df.columns)
        if retention_time_column
        else None
    )

    # col: charge_[1,2,3,...] = 0, 1
    charge_map = {
        col: int(re.search(r"(\d+)", col).group(1))
        for col in pin_df.columns
        if re.search(r"charge[_]?(\d+)", col, re.IGNORECASE)
    }

    def extract_charge(row):
        for col, num in charge_map.items():
            if int(float(row[col])) == 1:
                return num
        return None

    pin_df["Charge"] = pin_df.apply(extract_charge, axis=1)

    # feature columns: columns that are not non-feature columns
    non_feature_columns = [label, scan, specid, peptide, protein, hit_rank, "Charge"]
    feature_columns = [col for col in pin_df.columns if col not in non_feature_columns]

    logger.info(
        f"Columns: label={label}, scan={scan}, specid={specid}, peptide={peptide}, "
        f"protein={protein}, hit_rank={hit_rank}, retention_time={retention_time_column}, "
        f"features={feature_columns}"
    )

    pin_df[scan] = pin_df[scan].astype(int)
    pin_df[specid] = pin_df[specid].astype(str)
    pin_df[peptide] = pin_df[peptide].astype(str)
    pin_df[protein] = pin_df[protein].astype(str)
    pin_df[hit_rank] = pin_df[hit_rank].astype(float).astype(int)
    pin_df["Charge"] = pin_df["Charge"].astype(float).astype(int)
    if retention_time_column:
        pin_df[retention_time_column] = pin_df[retention_time_column].astype(float)
    for col in feature_columns:
        pin_df[col] = pin_df[col].astype(float)

    # label = 1 for target, -1 for decoy. Convert to Boolean.
    pin_df[label] = pin_df[label] == "1"
    rescoring_features = {"Original": feature_columns}

    return PsmContainer(
        psms=pin_df,
        label_column=label,
        scan_column=scan,
        spectrum_column=specid,
        ms_data_file_column=None,
        peptide_column=peptide,
        protein_column=protein,
        charge_column="Charge",
        rescoring_features=rescoring_features,
        hit_rank_column=hit_rank,
        retention_time_column=retention_time_column,
    )

_read_single_pin_as_df(pin_file)

Read a single PIN file into a DataFrame.

Parameters:

Name Type Description Default
pin_file str

The file path to the PIN file.

 required

Returns:

Type Description
DataFrame

A DataFrame containing the PSM data.
Notes

This function: 1. Reads the PIN file header 2. Processes the proteins column as a tab-separated list 3. Creates a DataFrame with the processed data

Source code in optimhc/parser/pin.py 
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
def _read_single_pin_as_df(pin_file: str) -> pd.DataFrame:
    """
    Read a single PIN file into a DataFrame.

    Parameters
    ----------
    pin_file : str
        The file path to the PIN file.

    Returns
    -------
    pd.DataFrame
        A DataFrame containing the PSM data.

    Notes
    -----
    This function:
    1. Reads the PIN file header
    2. Processes the proteins column as a tab-separated list
    3. Creates a DataFrame with the processed data
    """
    logger.info(f"Reading PIN file: {pin_file}")
    with open(pin_file, "r") as f:
        header = f.readline().strip().split("\t")
        header_len = len(header)
        data = []
        for line in f:
            parts = line.strip().split("\t")
            proteins_column_num = len(parts) - header_len + 1
            proteins = "\t".join(parts[-proteins_column_num:])
            data.append(parts[: len(parts) - proteins_column_num] + [proteins])
    df = pd.DataFrame(data, columns=header)
    logger.debug(f"Header: {header}")
    return df

mzML Parser

mzml

logger = logging.getLogger(__name__) module-attribute

extract_mzml_data(mzml_filename, scan_ids=None)

Extract scan data from an mzML file.

Parameters:

Name Type Description Default
mzml_filename str

The path to the mzML file.

 required
scan_ids list[int] or None

A list of scan IDs to extract. If None, extracts all scans.

 None

Returns:

Type Description
DataFrame

A DataFrame containing the extracted scan data with columns: - source: The source file name - scan: The scan ID - mz: The m/z values array - intensity: The intensity values array - charge: The charge state - retention_time: The retention time
Notes

This function: 1. Reads the mzML file using pyteomics 2. Extracts scan data including retention time, charge state, m/z values, and intensities 3. Filters scans based on provided scan IDs if specified 4. Returns a DataFrame with the extracted data

Source code in optimhc/parser/mzml.py 
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
def extract_mzml_data(mzml_filename, scan_ids=None):
    """
    Extract scan data from an mzML file.

    Parameters
    ----------
    mzml_filename : str
        The path to the mzML file.
    scan_ids : list[int] or None, optional
        A list of scan IDs to extract. If None, extracts all scans.

    Returns
    -------
    pd.DataFrame
        A DataFrame containing the extracted scan data with columns:
        - source: The source file name
        - scan: The scan ID
        - mz: The m/z values array
        - intensity: The intensity values array
        - charge: The charge state
        - retention_time: The retention time

    Notes
    -----
    This function:
    1. Reads the mzML file using pyteomics
    2. Extracts scan data including retention time, charge state, m/z values, and intensities
    3. Filters scans based on provided scan IDs if specified
    4. Returns a DataFrame with the extracted data
    """
    filename = mzml_filename.split("/")[-1].replace(".mzML", "")
    logger.info(f"Extracting scans from {mzml_filename}")

    scan_ids = set(scan_ids) if scan_ids is not None else None

    (
        extracted_scan_ids,
        mzml_filenames,
        intensities,
        mz_values,
        charges,
        retention_times,
    ) = ([], [], [], [], [], [])

    try:
        with mzml.read(mzml_filename) as reader:
            for spectrum in reader:
                try:
                    scan_id = int(spectrum["id"].split("scan=")[-1])

                    if scan_ids is not None and scan_id not in scan_ids:
                        continue

                    mz_array = np.array(spectrum.get("m/z array", []))
                    intensity_array = np.array(spectrum.get("intensity array", []))

                    charge = None
                    try:
                        charge = int(
                            spectrum["precursorList"]["precursor"][0]["selectedIonList"][
                                "selectedIon"
                            ][0]["charge state"]
                        )
                    except (KeyError, ValueError, IndexError):
                        pass

                    retention_time = None
                    try:
                        retention_time = float(spectrum["scanList"]["scan"][0]["scan start time"])
                    except (KeyError, ValueError, IndexError):
                        pass

                    extracted_scan_ids.append(scan_id)
                    mzml_filenames.append(filename)
                    intensities.append(intensity_array)
                    mz_values.append(mz_array)
                    charges.append(charge)
                    retention_times.append(retention_time)

                except Exception as e:
                    logger.warning(f"Skipping scan {scan_id} due to error: {e}")

    except Exception as e:
        logger.error(f"Failed to parse mzML file {mzml_filename}: {e}")
        raise RuntimeError(f"Error processing mzML file {mzml_filename}: {e}")

    data_dict = {
        "source": mzml_filenames,
        "scan": extracted_scan_ids,
        "mz": mz_values,
        "intensity": intensities,
        "charge": charges,
        "retention_time": retention_times,
    }

    scans_df = pd.DataFrame(data_dict)
    scans_df = scans_df.drop_duplicates(subset=["source", "scan"])

    logger.info(f"Successfully extracted {len(scans_df)} scans from {mzml_filename}")

    return scans_df