Dear GATK team, From what I read on the forum, it seems that HaplotypeCaller call variants (and perhaps calculates RGQ) using soft-clipped parts of the reads? Is there a parameter to suppress this behavior, making it throw out the soft-clipped proportions before calling? Best regards, Ray
I was softclipping some 84-bp-long single end reads using the following command:
java -Xmx2g -jar /GATK_Path/GenomeAnalysisTK.jar -T ClipReads -I imput_file.bam -o output_file.bam -R $ALIGNREF --cyclesToTrim "1-4,80-84" --clipRepresentation SOFTCLIP_BASES --fix_misencoded_quality_scores
There was no error while running, but when I looked at the file afterwards, every read had a cigar string of 84S. I'm fairly sure this is a bug. I was aligned to the human mitochondrial genome.
I am using GATK v2.3-9-ge5ebf34 on Mac OS 10.75
Hi, Some aligners produce Smith-Waterman alignments and may soft clip bases from a read when there are indels or mismatches near the ends of the reads. I was wondering if you include these bases in the realignment process? And if not whether you might consider making it an option? Thanks, Colin