2
$\begingroup$

I did RNA-seq of mammalian cells infected with pox virus. Now, I have read files which contain both host and virus reads. I want to align the reads both to host and viral genome. I was thinking I could concatenate the host and virus genome into one file and run salmon against this concatenated genome. However, salmon recommends transcriptome file for assembly which are not available for viruses. Virus genome are available as GenBank or GFF3 format from NCBI. Is there any way I can concatenate these formats into the format that can be used by salmon? Or is there any way around to use virus genome as reference in salmon?

$\endgroup$
  • $\begingroup$ To be clear, salmon does not align reads to the genome or transcriptome per se, but uses k-mers to determine which transcript from which a read is likely derived. If you really need the alignments, you'll need to use a different tool. But if you're just trying to do transcript quantification, tools like salmon or kallisto are MUCH faster and just as accurate as alignment-based tools. $\endgroup$ – Daniel Standage Mar 27 '19 at 13:12
2
$\begingroup$

You provide Salmon with a transcriptome fasta... so merging the human transcriptome with the pox virus genome fasta file should work. You don't mention that in your post but the viral genome should exist as a fasta. You can extract the sequences from genbank or maybe you can find our virus here: https://www.ebi.ac.uk/genomes/virus.html

You can find the coding transcriptome at the link below in the cdna folder. The non-coding transcriptome is in the ncrna. You should probably merge them:

ftp://ftp.ensembl.org/pub/release-95/fasta/homo_sapiens/

$\endgroup$
  • $\begingroup$ Yes, I have virus genome in fasta format. I was wondering if I just concatenate in plain fasta format , the features will be missed. How do I merge so that I will have virus features included in the merged file ? Thank you. $\endgroup$ – L R Joshi Mar 24 '19 at 19:18
  • $\begingroup$ If you want the virus features then you would need to find the ORF/CDS regions and subset your original fasta. You could try something like: ncbi.nlm.nih.gov/orffinder if they are not annotated. Alternatively they might already be present in Genbank/NCBI so you could just download them from there - although I'm not familiar enough with this to tell you how to do it in a high-throughput manner. $\endgroup$ – story Mar 25 '19 at 11:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.