Hard filtering to reduce false positive variants due to poly track sequences
In a old pipeline in my lab I have found this
#Annotate with low complexity region length using mdust
/share/apps/bcftools-distros/bcftools-1.3.1/bcftools annotate \
-a /state/partition1/db/human/gatk/2.8/b37/human_g1k_v37.mdust.v34.lpad1.bed.gz \
-c CHROM,FROM,TO,LCRLen \
-h <(echo '##INFO=<ID=LCRLen,Number=1,Type=Integer,Description="Overlapping mdust low complexity region length (mask cutoff: 34)">') \
-o "$seqId"_"$sampleId"_lcr.vcf \
"$seqId"_"$sampleId"_left_aligned_annotated.vcf
The link with the bed file is a old database GATK used to have to collect resources. Now, this is in here https://console.cloud.google.com/storage/browser/genomics-public-data/resources/broad/hg38/v0/ but I cannot find the equivalent for hg38.
This needs a licence https://genome.ucsc.edu/cgi-bin/hgTrackUi?g=rmsk ;(
1 answer
mdust low complexity region length
mdust no longer appears to be available but you could instead use sdust (https://github.com/lh3/sdust ) which is Heng Li's implementation. You will need to make up the BED file though.
Log in to answer this question.