博文
Overlapping variants called by GATK HaplotypeCaller
|
Recently, I found that overlapping varaints existed in the results that called by GATK HaplotypeCaller (version 4.0.1.2). I noticed this because I want to build consensus sequences using bcftools consensus, which produced warnings like these:
The site Chr01:597519 overlaps with another variant, skipping... The site Chr01:600176 overlaps with another variant, skipping... The site Chr01:914371 overlaps with another variant, skipping... ...
Then I checked the VCF file, which verified the weired variants existing.
Chr01 597511 . CTTCTTCTCTTTTTTTTT C ... Chr01 597519 . CT C ... Chr01 600174 . CTCTTT C ... Chr01 600176 . CT C ... Chr01 914369 . GTT G ... Chr01 914371 . T TGGGGG ...
The results were very strange, I don’t know why GATK HaplotypeCallerproduced these overlapping variants.
Furthermore, to update the correponding GFF/GTF files, creating CHAIN format file:
bcftools consensus -c out.chain -f ref.fasta -s sample_name -o out.fasta input.vcf.gz
Then convert the original GTF format file to new GTF file based the created consensus sequences using gtfToGenePred, liftOver, and genePredToGtf, which were part of the UCSC Genome Browser utilities.
gtfToGenePred old.gtf old.gp liftOver -genePred old.gp out.chain out.gp out.unMapped genePredToGtf file out.gp out.gtf
Another weired thing is that liftOver produced a bunch of unmapped genes (in out.unMapped).
https://blog.sciencenet.cn/blog-656335-1160082.html
上一篇：Picard MarkDuplicates running slower with more CPUs
下一篇：Extract unmapped read pairs from a bam file
