Haplotype from assembled reads

[amirshams84]

Hi, is there any way to generate haplotype from assembled reads using cliqueSNV

Thanks

[vtsyvina]

What do you mean? Just find consensus? There is an option for this, although even samtools can do it too. What kind or reads do you have, in what format, what haplotype do you expect to get? Some more info can help here

[amirshams84]

I have fastq reads from HiV study, my main goal is to generate haplotype and quasispecies I am following these route: A)standard 1) filter reads 2) map to reference 3) cliqueSNV(it works here :) )

B) assemble-based 1) filter reads 2)remove the host 3)assemble reads using spades 4) align to ref 5) detect haplotype or build a consensus sequence(does clique SNV works here??)

C) ref-free 1) filter reads 2) remove the host 3) detect haplotype(does cliqueSNV works here??)

[vtsyvina]

CliqueSNV works only with sam and bam. So you have to align reads one way or another. SO 2) I'm not familiar with spades, but if you align to ref then it should work. 3) no, we don't work reference-free, reads have to be aligned

[amirshams84]

Thanks for quick reply, I do appreciate it the issue is that since these assembled reads are low in number the read count for bam is very low like 3 reads total when I try cliqueSNV on this bam it gives me this error `CliqueSNV version: 1.5.3.3 Settings: {-m=snv-illumina, -in=try.bam} Reads number 3 SNV got 0 haplotypes

[] CliqueSNV didn't find any haplotypes (too low coverage) time,ms 383`

[vtsyvina]

Ah, I see now.

The only thing is then to find consensus (it is described in README for "-m" parameter). CliqueSNV won't be able to do anything else with such input.

[amirshams84]

generally, do you recommend to remove duplicate reads from bam using picard markduplicate or samtools rmdup

[vtsyvina]

I'm not sure about the nature of duplicate reads(didn't work closely with obtaining data from raw data), but I assume they may skew the frequencies of haplotypes if their distribution is not uniform. So they may be removed because of this reason

[amirshams84]

Thank you so much for the quick reply Another question is what is the best way to annotate this generated haplotype? my idea is degap generated haplotype, append with reference, and the multiple align and a dendrogram do you have a better idea?

[vtsyvina]

This question is out of my competence