needlestack icon indicating copy to clipboard operation
needlestack copied to clipboard

correct insertion extraction from mpileup

Open tdelhomme opened this issue 7 years ago • 1 comments

At the moment, samtools mpileup gives an "=" at the end of some insertions, and we have opened an issue for that.

This results in such a line in the VCF, were ALT contains (or is) an equal:

#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	IonXpress_005	IonXpress_019	IonXpress_020	IonXpress_030	IonXpress_053	IonXpress_104	IonXpress_105
chr11	62397115	.	-	=	152.793	.	TYPE=ins;NS=103;AF=0.00970874;DP=43980;RO=37749;AO=10;SRF=17260;SRR=20489;SAF=5;SAR=5;SOR=0.536285;RVSB=0.542758;FS=0;ERR=5.01679e-05;SIG=0.001;CONT=ATGxCCC;AC=0;AN=14	GT:QVAL:DP:RO:AO:AF:SB:SOR:RVSB:FS:QVAL_minAF:STATUS	0/0:0:202:177:0:0:112,65,0,0:-1:-1:0:1217.54:.	0/0:0:298:248:0:0:120,128,0,0:-1:-1:0:1850.99:.	0/0:0:306:244:0:0:99,145,0,0:-1:-1:0:1879.87:.	0/0:0:469:414:0:0:181,233,0,0:-1:-1:0:2952.58:.	0/0:0:401:349:0:0:138,211,0,0:-1:-1:0:2512.24:.	0/0:0:672:557:0:0:229,328,0,0:-1:-1:0:inf:.	0/0:0:713:623:0:0:322,301,0,0:-1:-1:0:inf:.

Either it is a bug from samtools and we just need to wait for their bug resolving or we need adapt the mpileup2readcounts script.

tdelhomme avatar May 09 '18 08:05 tdelhomme

This is actually an htslib 8yo bug reported here, for the moment we should adapt our mpileup2readcounts until they correct this.

tdelhomme avatar May 09 '18 09:05 tdelhomme