Remove invalid AAs in all pipelines (fixes #31)

ms2rescore/maxquant.py

@@ -48,6 +48,7 @@ def __init__(self, pandas_obj) -> None:
         """Pandas extension for MaxQuant msms.txt files."""
         self._obj = pandas_obj
         self._set_mass_error_unit()
+        self.invalid_amino_acids = r"[BJOUXZ]"
 
     @classmethod
     def from_file(
@@ -66,8 +67,8 @@ def from_file(
         filter_rank1_psms : bool, optional
             filter for rank 1 PSMs
         validate_amino_acids : bool, optional
-            remove PSMs where the sequence includes an invalid amino acid
-            (B, J, O, U, X, Z); required for MS2PIP compatibility
+            remove PSMs where the sequence includes an invalid amino acid; required for
+            MS2PIP compatibility
 
         Returns
         -------
@@ -114,7 +115,7 @@ def filter_rank1_psms(self) -> pd.DataFrame:
     def remove_invalid_amino_acids(self) -> pd.DataFrame:
         """Remove invalid amino acids from MSMS."""
         invalid_indices = self._obj[self._obj["Sequence"].str.contains(
-            r"[BJOUXZ]", regex=True
+            self.invalid_amino_acids, regex=True
         )].index
         self._obj = self._obj.drop(index=invalid_indices).reset_index(drop=True)
 

ms2rescore/peptideshaker.py

@@ -34,18 +34,33 @@ def __init__(self, pandas_obj: pd.DataFrame) -> None:
     def _validate(self):
         """Validate Pandas DataFrame as Extended PSM Report."""
         # TODO: Implement validation of PSM report DataFrame
-        pass
+        self.drop_invalid_amino_acids()
+
+    def drop_invalid_amino_acids(self, invalid_amino_acids=r"[BJOUXZ]"):
+        """Drop all PSMs (rows) with peptides containing invalid amino acids."""
+        to_drop = self._obj[
+            self._obj['Sequence'].str.contains(invalid_amino_acids, regex=True)
+        ].index
+        if len(to_drop) > 0:
+            logger.warning(
+                "Dropping %i PSMs from report due to invalid amino acids (%s)",
+                len(to_drop),
+                invalid_amino_acids
+            )
+            self._obj = self._obj.drop(index=to_drop)
 
     @staticmethod
     def from_tsv(path: Union[str, os.PathLike]) -> pd.DataFrame:
         """Read Extended PSM Report from TSV file."""
         ext_psm_report = pd.read_csv(path, sep="\t", index_col=0)
+        pd.ext_psm_report._validate(ext_psm_report)
         return ext_psm_report
 
     @staticmethod
     def from_xls(path: Union[str, os.PathLike]) -> pd.DataFrame:
         """Read Extended PSM Report from XLS file."""
         ext_psm_report = pd.read_excel(path, sheet_name=0, index_col=0)
+        pd.ext_psm_report._validate(ext_psm_report)
         return ext_psm_report
 
     @staticmethod

ms2rescore/percolator.py

@@ -38,6 +38,7 @@ def __init__(
         modification_mapping: Optional[
             Dict[Tuple[Union[str, None], Union[float, str]], str]
         ] = None,
+        invalid_amino_acids: Optional[str] = r"[BJOUXZ]"
     ):
         """
         Percolator In (PIN).
@@ -54,10 +55,14 @@ def __init__(
             (e.g. `{("Q", -17.02655): "Gln->pyro-Glu"}). If the keys are floats, they
             are rounded up to three decimals to avoid rounding issues while matching.
             If None, the original modification labels from the PIN file will be used.
+        invalid_amino_acids: str, optional
+            regex pattern of invalid amino acids. PSMs containing these amino acids will
+            be dropped. (default: `r"[BJOUXZ]"`)
 
         """
         # Attributes
         self.modification_pattern = r"\[([^\[^\]]*)\]"
+        self.invalid_amino_acids = invalid_amino_acids
 
         # Parameters
         self.path = path
@@ -301,6 +306,23 @@ def get_spectrum_filename(
         else:
             raise ValueError("Multiple spectrum filenames found in single PIN file.")
 
+    def drop_invalid_amino_acids(self):
+        """Drop all PSMs (rows) with peptides containing invalid amino acids."""
+        if "sequence" in self.df.columns:
+            sequences = self.df['sequence']
+        else:
+            sequences = self._get_sequence_column()
+        to_drop = sequences[
+            sequences.str.contains(self.invalid_amino_acids, regex=True)
+        ].index
+        if len(to_drop) > 0:
+            logger.warning(
+                "Dropping %i PSMs from PIN due to invalid amino acids (%s)",
+                len(to_drop),
+                self.invalid_amino_acids
+            )
+            self.df = self.df.drop(index=to_drop)
+
     @staticmethod
     def fix_tabs(
         path: str, id_column: str = "SpecId", prot_sep: Optional[str] = "|||"
@@ -376,6 +398,8 @@ def read(self, path: Optional[str] = None):
         if not self.path:
             raise ValueError("No path for PIN file defined.")
         self.df = pd.read_csv(self.fix_tabs(self.path), sep="\t")
+        if self.invalid_amino_acids:
+            self.drop_invalid_amino_acids()
 
     def write(self, path: Optional[str] = None):
         """Write PIN to file."""