bugfixes for errors in vcf with translocations and CIs

jbelyeu · Oct 20, 2020 · 3501f14 · 3501f14
1 parent b8e58fa
commit 3501f14
Show file tree

Hide file tree

Showing 4 changed files with 19 additions and 8 deletions.
diff --git a/samplot/samplot.py b/samplot/samplot.py
@@ -1712,6 +1712,8 @@ def plot_variant(sv, sv_type, ax, ranges):
     sv_title = ""
     if sv[0].chrm == sv[-1].chrm:
         sv_size = float(sv[0].end) - float(sv[0].start)
+        if len(sv) > 1:
+            sv_size = abs(int(float(sv[0].end) - float(sv[-1].start)))
         sv_size_unit = "bp"
 
         if sv_size > 1000000:

diff --git a/samplot/samplot_vcf.py b/samplot/samplot_vcf.py
@@ -441,11 +441,14 @@ def vcf(parser):
     out_file = open(args.command_file, "w")
 
     for var_count,variant in enumerate(vcf):
-        try:
-            translocation_chrom = variant.info.get("CHR2")
-        except:
-            translocation_chrom = None
+        translocation_chrom = None
         svtype = variant.info.get("SVTYPE", "SV")
+        if svtype in ["BND","TRA"]:
+            try:
+                translocation_chrom = variant.info.get("CHR2")
+            except:
+                pass
+
         if args.important_regions:
             if not var_in_important_regions(
                 important_regions, variant.chrom, variant.start, variant.stop
@@ -639,12 +642,12 @@ def vcf(parser):
 
         if "CIPOS" in variant.info:
             v = variant.info["CIPOS"]
-            cipos = "--start_ci '%s,%s'" % (abs(v[0]), abs(v[1]))
+            cipos = "--start_ci '%s,%s'" % (abs(int(v[0])), abs(int(v[1])))
         else:
             cipos = ""
         if "CIEND" in variant.info:
             v = variant.info["CIEND"]
-            ciend = "--end_ci '%s,%s'" % (abs(v[0]), abs(v[1]))
+            ciend = "--end_ci '%s,%s'" % (abs(int(v[0])), abs(int(v[1])))
         else:
             ciend = ""
         # dynamically set Z to speed drawing and remove noise for larger events
@@ -668,7 +671,7 @@ def vcf(parser):
             else:
                 title_list.append(variant_sample)
 
-        template = ("samplot plot {extra_args} -z {z} -n {titles}"
+        template = ("samplot plot {extra_args} -z {z} -n {titles} "
                 + "{cipos} {ciend} {svtype} -c {chrom} -s {start} "
                 + "-e {end} -o {fig_path} -d {downsample} -b {bams}")
         start = variant.start

diff --git a/samplot/templates/samplot_vcf.html b/samplot/templates/samplot_vcf.html
@@ -241,6 +241,9 @@ <h5>SV Overlaps</h5>
         let row = table.rows('.selected').data()[0]
         let img = new Image()
         img.src = `${row.svtype}_${row.chrom}_${row.start}_${row.end}.${plot_type}`
+        if ( row.chrom2) {
+            img.src = `${row.svtype}_${row.chrom}_${row.start}_${row.chrom2}_${row.end}.${plot_type}`
+        }
         let viewer = new Viewer(img, {
             hidden: function () {
                 viewer.destroy()

diff --git a/test/data/test.vcf b/test/data/test.vcf
@@ -1,6 +1,7 @@
 ##fileformat=VCFv4.1
 ##fileDate=20170929
 ##reference=ftp://ftp.1000genomes.ebi.ac.uk//vol1/ftp/technical/reference/phase2_reference_assembly_sequence/hs37d5.fa.gz
+##INFO=<ID=CHR2,Number=1,Type=String,Description="Chromosome for END coordinate">
 ##INFO=<ID=TSD,Number=1,Type=String,Description="Precise Target Site Duplication for bases, if unknown, value will be NULL">
 ##INFO=<ID=SVTYPE,Number=1,Type=String,Description="Type of structural variant">
 ##INFO=<ID=MSTART,Number=1,Type=Integer,Description="Mitochondrial start coordinate of inserted sequence">
@@ -73,7 +74,9 @@
 ##bcftools_viewCommand=view -c 1 -s NA12878 ALL.wgs.mergedSV.v8.20130502.svs.genotypes.vcf.gz
 ##bcftools_viewCommand=view -i '(SVTYPE="DEL" || SVTYPE="DUP" || SVTYPE="INV" || SVTYPE="INS")' test.vcf
 #CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	HG002	HG003	HG004
-1	24804398	1	T	<DEL>	2087.90	PASS	SVTYPE=DEL;CIEND=0,0;CIPOS=0,0;END=24807302;	GT:GQ:SQ:GL:DP:RO:AO:QR:QA:RS:AS:ASC:RP:AP:AB	0/1:146:581.24:-64,-6,-51:109:77:31:76:30:52:9:2:24:18:0.28	0/0:130:1506.66:-153,-15,-2:52:0:52:0:51:0:23:3:0:24:1	1/0:200:0.00:-0,-54,-181:182:182:0:181:0:120:0:0:61:0:0
+1	24804398	1	T	<DEL>	2087.90	PASS	SVTYPE=DEL;CIEND=100,100;CIPOS=1000,1000;END=24807302;	GT:GQ:SQ:GL:DP:RO:AO:QR:QA:RS:AS:ASC:RP:AP:AB	0/1:146:581.24:-64,-6,-51:109:77:31:76:30:52:9:2:24:18:0.28	0/0:130:1506.66:-153,-15,-2:52:0:52:0:51:0:23:3:0:24:1	1/0:200:0.00:-0,-54,-181:182:182:0:181:0:120:0:0:61:0:0
+1	24804398	1	T	<DEL>	2087.90	PASS	SVTYPE=TRA;CIEND=0,0;CIPOS=0,0;END=43059290;CHR2=1	GT:GQ:SQ:GL:DP:RO:AO:QR:QA:RS:AS:ASC:RP:AP:AB	0/1:146:581.24:-64,-6,-51:109:77:31:76:30:52:9:2:24:18:0.28	0/0:130:1506.66:-153,-15,-2:52:0:52:0:51:0:23:3:0:24:1	1/0:200:0.00:-0,-54,-181:182:182:0:181:0:120:0:0:61:0:0
+1	24804398	1	T	<DEL>	2087.90	PASS	SVTYPE=TRA;CIEND=0,0;CIPOS=0,0;END=99813787;CHR2=4	GT:GQ:SQ:GL:DP:RO:AO:QR:QA:RS:AS:ASC:RP:AP:AB	0/1:146:581.24:-64,-6,-51:109:77:31:76:30:52:9:2:24:18:0.28	0/0:130:1506.66:-153,-15,-2:52:0:52:0:51:0:23:3:0:24:1	1/0:200:0.00:-0,-54,-181:182:182:0:181:0:120:0:0:61:0:0
 1	43059290	2	T	<DEL>	2087.90	PASS	SVTYPE=DEL;CIEND=0,0;CIPOS=0,0;END=43059950;	GT:GQ:SQ:GL:DP:RO:AO:QR:QA:RS:AS:ASC:RP:AP:AB	0/0:146:581.24:-64,-6,-51:109:77:31:76:30:52:9:2:24:18:0.28	0/0:130:1506.66:-153,-15,-2:52:0:52:0:51:0:23:3:0:24:1	0/0:200:0.00:-0,-54,-181:182:182:0:181:0:120:0:0:61:0:0
 4	99813787	3	T	<DUP>	2087.90	PASS	SVTYPE=DUP;CIEND=0,0;CIPOS=0,0;END=99817098;	GT:GQ:SQ:GL:DP:RO:AO:QR:QA:RS:AS:ASC:RP:AP:AB	1/1:146:581.24:-64,-6,-51:109:77:31:76:30:52:9:2:24:18:0.28	1/1:130:1506.66:-153,-15,-2:52:0:52:0:51:0:23:3:0:24:1	1/1:200:0.00:-0,-54,-181:182:182:0:181:0:120:0:0:61:0:0
 11	67974432	4	T	<DUP>	2087.90	PASS	SVTYPE=DUP;CIEND=0,0;CIPOS=0,0;END=67975639;	GT:GQ:SQ:GL:DP:RO:AO:QR:QA:RS:AS:ASC:RP:AP:AB	0/1:146:581.24:-64,-6,-51:109:77:31:76:30:52:9:2:24:18:0.28	0/1:130:1506.66:-153,-15,-2:52:0:52:0:51:0:23:3:0:24:1	0/0:200:0.00:-0,-54,-181:182:182:0:181:0:120:0:0:61:0:0