Merge pull request #223 from hammerlab/maf_loading_fixes

Fixes to load_maf

test/data/tcga_ov.head.xychr.maf

@@ -0,0 +1,5 @@
+Hugo_Symbol	Entrez_Gene_Id	Center	NCBI_Build	Chromosome	Start_position	End_position	Strand	Variant_Classification	Variant_Type	Reference_Allele	Tumor_Seq_Allele1	Tumor_Seq_Allele2	dbSNP_RS	dbSNP_Val_Status	Tumor_Sample_Barcode	Matched_Norm_Sample_Barcode	Match_Norm_Seq_Allele1	Match_Norm_Seq_Allele2	Tumor_Validation_Allele1	Tumor_Validation_Allele2	Match_Norm_Validation_Allele1	Match_Norm_Validation_Allele2	Verification_Status	Validation_Status	Mutation_Status	Sequencing_Phase	Sequence_Source	Validation_Method	Score	BAM_file	Sequencer	Tumor_Sample_UUID	Matched_Norm_Sample_UUID	chromosome_name	start	stop	reference	variant	type	gene_name	transcript_name	transcript_species	transcript_source	transcript_version	strand	transcript_status	trv_type	c_position	amino_acid_change	ucsc_cons	domain	all_domains	deletion_substructures	transcript_error
+CDK11A	0	-	37	X	1650797	1650797	+	Missense_Mutation	SNP	A	A	G			TCGA-04-1337-01A-01W-0484-10	TCGA-04-1337-11A-01W-0485-10	A	A					Unknown	Unknown	Somatic	Phase_IV	Capture		1	dbGAP	-	1	1650797	1650797	A	G	SNP	CDK11A	ENST00000404249	human	ensembl	69_37n	-1	known	missense	c.325	p.C109R	0.971	NULL	pfam_Prot_kinase_cat_dom,pfam_Ser-Thr/Tyr_kinase_cat_dom,superfamily_Kinase-like_dom,smart_Ser/Thr_dual-sp_kinase_dom,smart_Tyr_kinase_cat_dom,pfscan_Prot_kinase_cat_dom	-	no_errors
+GNPAT	0	-	37	Y	231401797	231401797	+	Missense_Mutation	SNP	A	A	C			TCGA-04-1337-01A-01W-0484-10	TCGA-04-1337-11A-01W-0485-10	A	A					Unknown	Unknown	Somatic	Phase_IV	Capture		1	dbGAP	-	1	231401797	231401797	A	C	SNP	GNPAT	ENST00000366647	human	ensembl	69_37n	+1	known	missense	c.810	p.R270S	0.997	pfam_Acyltransferase,smart_Acyltransferase	pfam_Acyltransferase,smart_Acyltransferase	-	no_errors
+E2F2	0	-	37	1	23836447	23836447	+	Silent	SNP	C	C	A			TCGA-04-1337-01A-01W-0484-10	TCGA-04-1337-11A-01W-0485-10	C	C					Unknown	Unknown	Somatic	Phase_IV	Capture		1	dbGAP	-	1	23836447	23836447	C	A	SNP	E2F2	ENST00000361729	human	ensembl	69_37n	-1	known	silent	c.1239	p.L413	0.999	NULL	pfam_E2F_TDP	-	no_errors
+VSIG2	0	-	37	11	124617502	124617502	+	Missense_Mutation	SNP	C	C	G			TCGA-04-1337-01A-01W-0484-10	TCGA-04-1337-11A-01W-0485-10	C	C					Unknown	Unknown	Somatic	Phase_IV	Capture		1	dbGAP	-	11	124617502	124617502	C	G	SNP	VSIG2	ENST00000326621	human	ensembl	69_37n	-1	known	missense	c.913	p.G305R	0.813	NULL	pfam_Ig_V-set,pfam_Ig_I-set,pfam_Immunoglobulin,smart_Ig_sub,smart_Ig_sub2,smart_Ig_V-set_subgr,pfscan_Ig-like	-	no_errors

test/test_maf.py

@@ -96,3 +96,25 @@ def test_load_maf_dataframe():
             data_path("ov.wustle.subset5.maf"), raise_on_error=raise_on_error)
         eq_(len(variants_df), 5)
 
+
+def test_xy_contigs():
+    """
+    Test MAFs with X and Y chromosomes rather than just numerical chromosomes.
+    """
+    for raise_on_error in [True, False]:
+        variants = load_maf(
+            data_path("tcga_ov.head.xychr.maf"), raise_on_error=True)
+        eq_(len(variants), 4)
+
+
+def test_load_utf8():
+    """
+    Test MAFs loaded with utf-8 encoding.
+    """
+    for raise_on_error in [True, False]:
+        variants = load_maf(
+            data_path("ov.wustle.subset5.maf"), raise_on_error=True, encoding="utf-8")
+        eq_(len(variants), 5)
+        # Make sure we avoid "TypeError: character mapping must return integer, None or unicode"
+        # from Bio.Seq.
+        _ = variants.effects()

varcode/maf.py

@@ -17,7 +17,7 @@
 
 import pandas
 from typechecks import require_string
-from numpy import isnan
+from pandas import isnull
 
 from .reference import infer_genome
 from .variant import Variant, variant_ascending_position_sort_key
@@ -48,7 +48,7 @@
 ]
 
 
-def load_maf_dataframe(path, nrows=None, raise_on_error=True):
+def load_maf_dataframe(path, nrows=None, raise_on_error=True, encoding=None):
     """
     Load the guaranteed columns of a TCGA MAF file into a DataFrame
 
@@ -62,6 +62,9 @@ def load_maf_dataframe(path, nrows=None, raise_on_error=True):
 
     raise_on_error : bool
         Raise an exception upon encountering an error or log an error
+
+    encoding : str, optional
+        Encoding to use for UTF when reading MAF file.
     """
     require_string(path, "Path to MAF")
 
@@ -75,7 +78,8 @@ def load_maf_dataframe(path, nrows=None, raise_on_error=True):
         sep="\t",
         low_memory=False,
         skip_blank_lines=True,
-        header=0)
+        header=0,
+        encoding=encoding)
 
     if len(df.columns) < n_basic_columns:
         error_message = (
@@ -112,7 +116,8 @@ def load_maf(
         optional_cols=[],
         sort_key=variant_ascending_position_sort_key,
         distinct=True,
-        raise_on_error=True):
+        raise_on_error=True,
+        encoding=None):
     """
     Load reference name and Variant objects from MAF filename.
 
@@ -121,7 +126,7 @@ def load_maf(
 
     path : str
         Path to MAF (*.maf).
-      
+
     optional_cols : list, optional
         A list of MAF columns to include as metadata if they are present in the MAF.
         Does not result in an error if those columns are not present.
@@ -135,10 +140,13 @@ def load_maf(
 
     raise_on_error : bool
         Raise an exception upon encountering an error or just log a warning.
+
+    encoding : str, optional
+        Encoding to use for UTF when reading MAF file.
     """
     # pylint: disable=no-member
     # pylint gets confused by read_csv inside load_maf_dataframe
-    maf_df = load_maf_dataframe(path, raise_on_error=raise_on_error)
+    maf_df = load_maf_dataframe(path, raise_on_error=raise_on_error, encoding=encoding)
 
     if len(maf_df) == 0 and raise_on_error:
         raise ValueError("Empty MAF file %s" % path)
@@ -148,7 +156,7 @@ def load_maf(
     metadata = {}
     for _, x in maf_df.iterrows():
         contig = x.Chromosome
-        if not contig or isnan(contig):
+        if isnull(contig):
             error_message = "Invalid contig name: %s" % (contig,)
             if raise_on_error:
                 raise ValueError(error_message)
@@ -192,8 +200,8 @@ def load_maf(
         variant = Variant(
             contig,
             start_pos,
-            ref,
-            alt,
+            str(ref),
+            str(alt),
             ensembl=ensembl)
 
         # keep metadata about the variant and its TCGA annotation