Entering edit mode
5.0 years ago
snishtala03
▴
70
Hello,
I am working with RNA-Seq data for a virus. My collaborator gave me a predicted single contig reference which belongs to a particular genotype (let's say GenotypeA) of the virus. Now, some of my samples are from a different genotype (B,C or D) and therefore their mapping % when aligned to my predicted reference is very low (~24%) when I use STAR.
I think doing a de-novo assembly and then aligning the reads to the new reference should help me. Are there any tools which can do single contig de novo assembly? I tried Trinity but I did not get a single contig from it.
Thanks!
Just a thought: are you certain that all of your RNA-Seq data is pure virus? Have you checked that your other contigs are not something else, e.g. host (depending on how the virus was isolated, obviously)?
I am asking, as I have worked on RNA viruses present in insect transcriptomes. In my case, both Trinity and Velvet/Oases were able to assemble (nearly) full-length viruses among tons of other transcripts (e.g. from the host).
Ah, interesting! My host is human. I will try to remove human reads and then try this again.
You can also check the contigs/transcripts that you already have.
Appart from the ideas mentiond by @cschu181 I would not necessarily expect a single contig - there migth very well be regions witch are very hard to assemble due to repetitive sequences etc - such regions will cause the assemblies to be broken into many smaller fragments.