--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/gemini_mafify.py	Tue Mar 10 06:14:22 2020 -0400
@@ -0,0 +1,270 @@
+import string
+import sys
+
+
+so_to_maf = {
+    'splice_acceptor_variant': 'Splice_Site',
+    'splice_donor_variant': 'Splice_Site',
+    'transcript_ablation': 'Splice_Site',
+    'exon_loss_variant': 'Splice_Site',
+    'stop_gained': 'Nonsense_Mutation',
+    'stop_lost': 'Nonstop_Mutation',
+    'frameshift_variant': 'Frame_Shift_',
+    'initiator_codon_variant': 'Translation_Start_Site',
+    'start_lost': 'Translation_Start_Site',
+    'inframe_insertion': 'In_Frame_Ins',
+    'inframe_deletion': 'In_Frame_Del',
+    'conservative_inframe_insertion': 'In_Frame_Ins',
+    'conservative_inframe_deletion': 'In_Frame_Del',
+    'disruptive_inframe_insertion': 'In_Frame_Ins',
+    'disruptive_inframe_deletion': 'In_Frame_Del',
+    'missense_variant': 'Missense_Mutation',
+    'coding_sequence_variant': 'Missense_Mutation',
+    'conservative_missense_variant': 'Missense_Mutation',
+    'rare_amino_acid_variant': 'Missense_Mutation',
+    'transcript_amplification': 'Intron',
+    'intron_variant': 'Intron',
+    'INTRAGENIC': 'Intron',
+    'intragenic_variant': 'Intron',
+    'splice_region_variant': 'Splice_Region',
+    'mature_miRNA_variant': 'RNA',
+    'exon_variant': 'RNA',
+    'non_coding_exon_variant': 'RNA',
+    'non_coding_transcript_exon_variant': 'RNA',
+    'non_coding_transcript_variant': 'RNA',
+    'nc_transcript_variant': 'RNA',
+    'stop_retained_variant': 'Silent',
+    'synonymous_variant': 'Silent',
+    'NMD_transcript_variant': 'Silent',
+    'incomplete_terminal_codon_variant': 'Silent',
+    '5_prime_UTR_variant': "5'UTR",
+    '5_prime_UTR_premature_start_codon_gain_variant': "5'UTR",
+    '3_prime_UTR_variant': "3'UTR",
+    'intergenic_variant': 'IGR',
+    'intergenic_region': 'IGR',
+    'regulatory_region_variant': 'IGR',
+    'regulatory_region': 'IGR',
+    'TF_binding_site_variant': 'IGR',
+    'upstream_gene_variant': "5'Flank",
+    'downstream_gene_variant': "3'Flank",
+}
+
+
+class VariantEffect():
+    def __init__(self, variant_type):
+        self.variant_type = variant_type.capitalize()
+        assert self.variant_type in ['Snp', 'Ins', 'Del']
+
+    def __getitem__(self, so_effect):
+        if so_effect not in so_to_maf or (
+            'frame' in so_effect and self.variant_type == 'Snp'
+        ):
+            return 'Targeted_Region'
+
+        ret = so_to_maf[so_effect]
+        if ret == 'Frame_Shift_':
+            ret += self.variant_type
+        return ret
+
+
+infile = sys.argv[1]
+if len(sys.argv) > 2:
+    tumor_sample_name = sys.argv[2]
+if len(sys.argv) > 3:
+    normal_sample_name = sys.argv[3]
+
+start_pos_idx = None
+ref_idx = None
+alt_idx = None
+variant_type_idx = None
+variant_classification_idx = None
+gt_alt_depths_idx = {}
+gt_ref_depths_idx = {}
+gts_idx = {}
+samples = set()
+required_fields = [
+    'Hugo_Symbol',
+    'NCBI_Build',
+    'Variant_Type',
+    'Variant_Classification',
+    'Tumor_Sample_Barcode',
+    'HGVSp_Short'
+]
+
+
+with open(infile) as data_in:
+    cols = data_in.readline().rstrip().split('\t')
+    for field in required_fields:
+        if field not in cols:
+            raise IndexError(
+                'Cannot generate valid MAF without the following input '
+                'columns: {0}.\n'
+                'Missing column: "{1}"'
+                .format(required_fields, field)
+            )
+    for i, col in enumerate(cols):
+        if col == 'Variant_Type':
+            variant_type_idx = i
+        elif col == 'Variant_Classification':
+            variant_classification_idx = i
+        elif col == 'Start_Position':
+            start_pos_idx = i
+        elif col == 'Reference_Allele':
+            ref_idx = i
+        elif col == 'alt':
+            alt_idx = i
+        else:
+            column, _, sample = col.partition('.')
+            if sample:
+                if column == 'gt_alt_depths':
+                    gt_alt_depths_idx[sample] = i
+                elif column == 'gt_ref_depths':
+                    gt_ref_depths_idx[sample] = i
+                elif column == 'gts':
+                    gts_idx[sample] = i
+                else:
+                    # not a recognized sample-specific column
+                    continue
+                samples.add(sample)
+
+    if ref_idx is None:
+        raise IndexError('Input file does not have a column "Reference_Allele".')
+
+    if not tumor_sample_name:
+        if normal_sample_name:
+            raise ValueError(
+                'Normal sample name requires the tumor sample name to be '
+                'specified, too.'
+            )
+        if len(samples) > 1:
+            raise ValueError(
+                'A tumor sample name is required with more than one sample '
+                'in the input.'
+            )
+        if samples:
+            # There is a single sample with genotype data.
+            # Assume its the tumor sample.
+            tumor_sample_name = next(iter(samples))
+    else:
+        if tumor_sample_name not in samples:
+            raise ValueError(
+                'Could not find information about the specified tumor sample '
+                'in the input.'
+            )
+        if tumor_sample_name == normal_sample_name:
+            raise ValueError(
+                'Need different names for the normal and the tumor sample.'
+            )
+
+    if normal_sample_name and normal_sample_name not in samples:
+        raise ValueError(
+            'Could not find information about the specified normal sample '
+            'in the input.'
+        )
+
+    # All input data checks passed!
+    # Now extract just the relevant index numbers for the tumor/normal pair
+    gts_idx = (
+        gts_idx.get(tumor_sample_name, alt_idx),
+        gts_idx.get(normal_sample_name)
+    )
+    gt_alt_depths_idx = (
+        gt_alt_depths_idx.get(tumor_sample_name),
+        gt_alt_depths_idx.get(normal_sample_name)
+    )
+    gt_ref_depths_idx = (
+        gt_ref_depths_idx.get(tumor_sample_name),
+        gt_ref_depths_idx.get(normal_sample_name)
+    )
+
+    # Echo all MAF column names
+    cols_to_print = []
+    for n in range(len(cols)):
+        if n in gts_idx:
+            continue
+        if n in gt_alt_depths_idx:
+            continue
+        if n in gt_ref_depths_idx:
+            continue
+        if n != alt_idx:
+            cols_to_print.append(n)
+
+    print('\t'.join([cols[n] for n in cols_to_print]))
+
+    for line in data_in:
+        cols = line.rstrip().split('\t')
+
+        gt_alt_depths = [
+            int(cols[ad_idx]) if ad_idx else ''
+            for ad_idx in gt_alt_depths_idx
+        ]
+        gt_ref_depths = [
+            int(cols[rd_idx]) if rd_idx else ''
+            for rd_idx in gt_ref_depths_idx
+        ]
+
+        gts = [
+            ['', ''],
+            ['', '']
+        ]
+        for n, gt_idx in enumerate(gts_idx):
+            if gt_idx:
+                gt_sep = '/' if '/' in cols[gt_idx] else '|'
+                allele1, _, allele2 = [
+                    '' if allele == '.' else allele
+                    for allele in cols[gt_idx].partition(gt_sep)
+                ]
+                # follow cBioportal recommendation to leave allele1 empty
+                # when information is not avaliable
+                if not allele2:
+                    gts[n] = [allele2, allele1]
+                else:
+                    gts[n] = [allele1, allele2]
+        if not gts:
+            gts = [['', ''], ['', '']]
+
+        if cols[variant_type_idx].lower() in ['ins', 'del']:
+            # transform VCF-style indel representations into MAF ones
+            ref_allele = cols[ref_idx]
+            for n, nucs in enumerate(
+                zip(
+                    ref_allele,
+                    *[allele for gt in gts for allele in gt if allele]
+                )
+            ):
+                if any(nuc != nucs[0] for nuc in nucs[1:]):
+                    break
+            else:
+                n += 1
+            if n > 0:
+                cols[ref_idx] = cols[ref_idx][n:] or '-'
+                for gt in gts:
+                    for idx, allele in enumerate(gt):
+                        if allele:
+                            gt[idx] = allele[n:] or '-'
+                if cols[ref_idx] == '-':
+                    n -= 1
+                cols[start_pos_idx] = str(int(cols[start_pos_idx]) + n)
+
+        # in-place substitution of so_effect with MAF effect
+        cols[variant_classification_idx] = VariantEffect(
+            cols[variant_type_idx]
+        )[cols[variant_classification_idx]]
+        ret_line = '\t'.join([cols[n] for n in cols_to_print])
+
+        field_formatters = {
+            'tumor_seq_allele1': gts[0][0],
+            'tumor_seq_allele2': gts[0][1],
+            'match_norm_seq_allele1': gts[1][0],
+            'match_norm_seq_allele2': gts[1][1],
+            't_alt_count': gt_alt_depths[0],
+            'n_alt_count': gt_alt_depths[1],
+            't_ref_count': gt_ref_depths[0],
+            'n_ref_count': gt_ref_depths[1],
+        }
+
+        print(
+            # use safe_substitute here to avoid key errors with column content
+            # looking like unknown placeholders
+            string.Template(ret_line).safe_substitute(field_formatters)
+        )