Allow annotation of Number=A,R INFO tags from tab-delimited files when alleles are split across multiple records

Author: pd3

NEWS

@@ -11,6 +11,9 @@ Changes affecting specific commands:
     - Fix a bug in annotation renaming via `--rename-annots` and `-c NEW:=OLD`,
       where renaming INFO/TAG would simultaneously rename also FORMAT/TAG (#2489)
 
+    - Allow annotation of Number=A,R INFO tags from tab-delimited files when alleles
+      are split across multiple records.
+
 * bcftools cnv
 
     - Fix off-by-one error in reporting HMM qualities

test/annotate.AR.1.out

@@ -0,0 +1,9 @@
+##fileformat=VCFv4.2
+##FILTER=<ID=PASS,Description="All filters passed">
+##contig=<ID=1>
+##INFO=<ID=IR,Number=R,Type=Integer,Description="">
+##INFO=<ID=FR,Number=R,Type=Float,Description="">
+##INFO=<ID=IA,Number=A,Type=Integer,Description="">
+##INFO=<ID=FA,Number=A,Type=Float,Description="">
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO
+1	100	.	A	C,G	.	.	IA=10,20;FA=1.1,2.2;IR=11,22,44;FR=1.1,2.2,4.4

test/annotate.AR.2.out

@@ -0,0 +1,9 @@
+##fileformat=VCFv4.2
+##FILTER=<ID=PASS,Description="All filters passed">
+##contig=<ID=1>
+##INFO=<ID=IR,Number=R,Type=Integer,Description="">
+##INFO=<ID=FR,Number=R,Type=Float,Description="">
+##INFO=<ID=IA,Number=A,Type=Integer,Description="">
+##INFO=<ID=FA,Number=A,Type=Float,Description="">
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO
+1	100	.	A	C,G	.	.	IA=10,20;FA=1.1,2.2;IR=33,22,44;FR=3.3,2.2,4.4

test/annotate.AR.tab

@@ -0,0 +1,2 @@
+1	100	A	C	10	1.1	11,22	1.1,2.2
+1	100	A	G	20	2.2	33,44	3.3,4.4

test/annotate.AR.vcf

@@ -0,0 +1,9 @@
+##fileformat=VCFv4.2
+##FILTER=<ID=PASS,Description="All filters passed">
+##contig=<ID=1>
+##INFO=<ID=IR,Number=R,Type=Integer,Description="">
+##INFO=<ID=FR,Number=R,Type=Float,Description="">
+##INFO=<ID=IA,Number=A,Type=Integer,Description="">
+##INFO=<ID=FA,Number=A,Type=Float,Description="">
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO
+1	100	.	A	C,G	.	.	.

test/test.pl

@@ -565,7 +565,9 @@
 run_test(\&test_vcf_sort,$opts,in=>'sort',out=>'sort.out',args=>q[-m 0],fmt=>'%CHROM\\t%POS\\t%REF,%ALT\\n');
 run_test(\&test_vcf_sort,$opts,in=>'sort',out=>'sort.out',args=>q[-m 1000],fmt=>'%CHROM\\t%POS\\t%REF,%ALT\\n');
 run_test(\&test_vcf_regions,$opts,in=>'regions');
-+run_test(\&test_vcf_annotate,$opts,in=>'annotate.int.dst',vcf=>'annotate.int.src',out=>'annotate.int.1.out',args=>q[-c +INFO/IA,+INFO/IR]);
+run_test(\&test_vcf_annotate,$opts,in=>'annotate.AR',tab=>'annotate.AR',out=>'annotate.AR.1.out',args=>q[-c CHROM,POS,REF,ALT,+INFO/IA,+INFO/FA,+INFO/IR,+INFO/FR]);
+run_test(\&test_vcf_annotate,$opts,in=>'annotate.AR',tab=>'annotate.AR',out=>'annotate.AR.2.out',args=>q[-c CHROM,POS,REF,ALT,INFO/IA,INFO/FA,INFO/IR,INFO/FR]);
+run_test(\&test_vcf_annotate,$opts,in=>'annotate.int.dst',vcf=>'annotate.int.src',out=>'annotate.int.1.out',args=>q[-c +INFO/IA,+INFO/IR]);
 run_test(\&test_vcf_annotate,$opts,in=>'annotate.rename',out=>'annotate.rename.1.out',args=>q[-c INFO/ii:=INFO/XX]);
 run_test(\&test_vcf_annotate,$opts,in=>'annotate.rename',out=>'annotate.rename.2.out',args=>q[-c FORMAT/ff:=FORMAT/XX]);
 run_test(\&test_vcf_annotate,$opts,in=>'annotate36',bed=>'annots36',out=>'annotate36.1.out',args=>q[-c CHROM,POS,-,~X,-,AF,-,-,- -i'SVTYPE={X}' -k]);

vcfannotate.c

@@ -810,7 +810,8 @@ static int setter_ARinfo_int32(args_t *args, bcf1_t *line, annot_col_t *col, int
 
         args->tmpi2[i] = args->tmpi[ map[i] ];
     }
-    return bcf_update_info_int32(args->hdr_out,line,col->hdr_key_dst,args->tmpi2,ndst);
+    int ret = bcf_update_info_int32(args->hdr_out,line,col->hdr_key_dst,args->tmpi2,ndst);
+    return ret<0 ? ret : 1;
 }
 static int setter_info_int(args_t *args, bcf1_t *line, annot_col_t *col, void *data)
 {
@@ -919,7 +920,10 @@ static int setter_info_int(args_t *args, bcf1_t *line, annot_col_t *col, void *d
     }
 
     if ( col->number==BCF_VL_A || col->number==BCF_VL_R )
+    {
+        assert(tab);
         return setter_ARinfo_int32(args,line,col,tab->nals,tab->als,ntmpi);
+    }
 
     if ( col->replace & REPLACE_MISSING )
     {
@@ -975,7 +979,8 @@ static int setter_ARinfo_real(args_t *args, bcf1_t *line, annot_col_t *col, int
 
         args->tmpf2[i] = args->tmpf[ map[i] ];
     }
-    return bcf_update_info_float(args->hdr_out,line,col->hdr_key_dst,args->tmpf2,ndst);
+    int ret = bcf_update_info_float(args->hdr_out,line,col->hdr_key_dst,args->tmpf2,ndst);
+    return ret<0 ? ret : 1;
 }
 static int setter_info_real(args_t *args, bcf1_t *line, annot_col_t *col, void *data)
 {
@@ -1103,7 +1108,10 @@ static int setter_info_real(args_t *args, bcf1_t *line, annot_col_t *col, void *
     }
 
     if ( col->number==BCF_VL_A || col->number==BCF_VL_R )
+    {
+        assert(tab);
         return setter_ARinfo_real(args,line,col,tab->nals,tab->als,ntmpf);
+    }
 
     if ( col->replace & REPLACE_MISSING )
     {
@@ -3789,7 +3797,7 @@ static int annotate_from_vcf(args_t *args, bcf1_t *line)
     for (j=0; j<args->ncols; j++)
     {
         if ( !args->cols[j].setter ) continue;
-        if ( args->cols[j].setter(args,line,&args->cols[j],aline) )
+        if ( args->cols[j].setter(args,line,&args->cols[j],aline) < 0 )
             error("fixme: Could not set %s at %s:%"PRId64"\n", args->cols[j].hdr_key_src,bcf_seqname(args->hdr,line),(int64_t) line->pos+1);
     }
     return 1;
@@ -3800,7 +3808,7 @@ static int annotate_from_self(args_t *args, bcf1_t *line)
     for (j=0; j<args->ncols; j++)
     {
         if ( !args->cols[j].setter ) continue;
-        if ( args->cols[j].setter(args,line,&args->cols[j],NULL) )
+        if ( args->cols[j].setter(args,line,&args->cols[j],NULL) < 0 )
             error("fixme: Could not set %s at %s:%"PRId64"\n", args->cols[j].hdr_key_src,bcf_seqname(args->hdr,line),(int64_t) line->pos+1);
     }
     return 0;