pymaf.py

"""
Helper functions to work with MAF files. Based on maftools for R.

Author: Matthew Nguyen, 2019
"""

import pandas as pd


"""
Class storing summarizations of MAF files.
"""
class MAF():

    def __init__(self, maf, vc=None, rdup=True):
        if vc:
            self.vc_nonSyn = vc
        else:
            self.vc_nonSyn = {"Frame_Shift_Del", "Frame_Shift_Ins", "Splice_Site", "Translation_Start_Site",
                              "Nonsense_Mutation", "Nonstop_Mutation", "In_Frame_Del", "In_Frame_Ins",
                              "Missense_Mutation"}
        self.rdup = rdup
        self.nonSyn_df, self.syn_df = self.read_maf(maf)
        self.variant_count = self.variants_per_sample(self.nonSyn_df)
        self.codingVars = set(self.nonSyn_df['Variant_Classification'].unique())

    """
    Reads MAF file, and returns nonsynonymous variants in a dataframe.
    """
    def read_maf(self, maf):
        print("Reading MAF...")
        maf_df = pd.read_csv(maf, sep='\t', dtype={'Chromosome': str})

        # Remove synonymous variants
        nonSyn_df = maf_df[maf_df['Variant_Classification'].isin(self.vc_nonSyn)]
        syn_df = maf_df[~maf_df['Variant_Classification'].isin(self.vc_nonSyn)]

        if self.rdup:
            duplicate_rows = nonSyn_df.duplicated(['Chromosome', 'Start_Position', 'Tumor_Sample_Barcode'])
            nonSyn_df = nonSyn_df.loc[~duplicate_rows]
            print("--Removed {0} duplicated variants--".format(duplicate_rows.sum()))


        return nonSyn_df, syn_df

    """
    Returns dataframe with number of variants per sample.
    """
    def variants_per_sample(self, maf_df):
        variant_count_series = maf_df.groupby('Tumor_Sample_Barcode').size().sort_values(ascending=False)
        variant_count_series = variant_count_series.rename_axis(None)
        variant_count_df = variant_count_series.reset_index(level=0)
        variant_count_df.columns = ['Tumor_Sample_Barcode', 'Variants']
        return variant_count_df