large deletion detection in a haploid microbial eukaryote
2
0
Entering edit mode
4.9 years ago
thomas.welch ▴ 50

Hi there, this is a very general question i'm afraid. but i've been searching for some time for a solution and can't find i concrete answer.

I will soon have approximately 180 paired-end sequenced genome (x8-x13 coverage) of a haploid microbial genome, for which i am putting together a pipeline for genotyping of SNPs and analysis for a GWAS. However i would also like to (in a seperate pipeline and analysis) genotype for large deletions, and presence/absence of supernumerary chromosome present in this organism. I've been looking around for tools capable of this but most of them seem tuned toward human genomes (I recently looked at the possibility of using genome STRiP and lumpy). another possibility i suppose could also be to infer structural variants from copy number variants using CNVnator, but i don't think i have enough coverage to justify this approach.

Is there any way that i can genotype these structural variants?

P.S. this is my first time working with sequence data, so i may be a little slow on the uptake, any help would be greatly appreciated.

structural variant haploid microbial sequencing • 948 views
1
Entering edit mode
4.9 years ago
Macspider ★ 3.4k

Pindel is what you are searching for: http://gmt.genome.wustl.edu/packages/pindel/

Be careful, it is slow as hell, so you might want to split your bam file into as many chunks as there are cores in your cluster.

0
Entering edit mode
4.9 years ago

If you map with BBMap, which is quite fast, long deletions will be detected within the reads. You can add the flag "maxindel=200000" or similar to increase the detection length (default is only 16000). Subsequently, you can call them with BBMap's CallVariants tool, which is also very fast. Specifically:

bbmap.sh in=reads.fq ref=microbe.fa out=mapped.sam maxindel=200000
callvariants.sh in=mapped.sam ref=microbe.fa out=vars.vcf


You can specify a ploidy for CallVariants, but it defaults to 1 so you don't need to in this case.