Hi all,
We have multiple sequence alignments from thousands of genes and we are wondering what to use to index all the point mutations/insertion/deletion sites without having to manually inspect the alignments. The index has to point out the location of the mutations and what strains they belong to. Is there a free software/website that can do this?
Thanks everyone.
Hello,
The difficulty in what you're asking is the fact that most variant callers do not care in which sequence (sequencing read) the variation is located, only on which position. Thus, you would get the positions, but not the strains, and you will have to manually go to that position and look those up. I can think of kind of solution, but it will require some manual work too.
I am assuming here that the number of strains you have is much less than the number of genes. I also assume you have a single sequence for each strain (e.g. gene sequences downloaded from a database). What you can do is look for a way to automate a variant caller like samtools or a Galaxy pipeline and run each strain by itself. You may need to multiply the sequences so that you simulate coverage higher than 1x. You will effectively get "homozygous" calls for all variants that are present in the strain.
Putting all the strains together may be a bit of a hassle, again depending on their number. Since callers usually output VCF files you could use VCFtools (vcftools.sourceforge.net) to compare strains and merge files, but you should take care not to lose information, since VCFtools again assumes you have sequencing data and not an MSA from single "coverage" strains. An excel table may do the job if the strains are not too big a number.
Or you could find someone to write a script for combining the resultant VCF files into by-strain or by-variant statistics to your liking.
Hi,
You should do the variant calling with all strains at the same time, because the quality gets much better. Variants may have little support (only a few reads) in one strain, but if the same variant occurs in some other strains as well it is probably real. This information is used by several tools during variant detection.
Samtools and bcftools could do the variant detection if you have the multiple sequence alignment in BAM format:
http://www.htslib.org/doc/samtools.html
https://samtools.github.io/bcftools/bcftools.html#call
There is an example for the variant detection at the bottom of the samtools website.
Good luck!
- Badges
- Science topic
- Similar topics
- Filing