Forgive the basic question here, but I'm super novice at this ...

I have a series of paired-end RNA fastq files (e.g. SRR10720226_1.fastq.gz and SRR10720226_2.fastq.gz). I'm trying to figure out what software I need to use to translate those into proteins.

I was trying to use biopython to do this using the translate method. The problem is biopython doesn't seem to consider both files together, only one fastq at a time.

Which tool should I use to work with the reads together? Is there anything else I should be doing besides just translating to protein?


1 Answer 1


This is a much more complicated question than it might seem.

First, you need to understand how RNA-Seq works, and what your data really is. Your "paired-end files" contain reads, which will contain fragments of transcripts.

Since those are only fragments of transcripts, you don't know what frame they are in, so you don't have a single possible translation. And each read would only give you a small part of the protein.

Since the data is for humans, your best course of action would be to first, align the reads to the annotated human genome (or transcriptome). Then, count the number of reads for each gene (or transcript), and decide on a condition to call a gene (or transcript) present in the sample (typically by enrichment between two sets of samples, in your case a threshold might be good enough). Finally, for the genes (or transcripts) present in the sample, find out what protein (or isoform) they correspond to, and if needed retrieve the sequence.

To be clear, you will have to learn a lot more to get that done. For the first step, easier tools might include kallisto or salmon. You would probably want to use biological replicates; SRP237926 seems to have several samples, reading the paper would be necessary to determine whether they are useful.

  • $\begingroup$ Thanks for the suggestions. It's a little unclear to me exactly what kallisto and salmon are doing. Are they doing alignment like bwa, or something else? What's the result out of those programs? It doesn't appear to be a list of proteins. Sorry again if I'm saying things that don't make sense. Still trying to get my feet under me here. $\endgroup$
    – agf1997
    Apr 28, 2022 at 5:02
  • 1
    $\begingroup$ Yes, like bwa, they will let you determine what gene a read comes from (after bwa, you would still need to count the reads within genomic loci). But they actually use a totally different approach that gives basically the same result, and are easier to run. I would recommend to read the papers, although this blog post about kallisto gives a nice background. $\endgroup$
    – Alexlok
    Apr 28, 2022 at 5:23
  • $\begingroup$ Any papers you could suggest would be helpful. Something of a 101 on the subject would be nice. Seems like there's lots on info out there but it all assumes a certain level of knowledge that I don't have. $\endgroup$
    – agf1997
    Apr 28, 2022 at 16:24
  • $\begingroup$ For instance, I'm not sure how kallisto fits into the pipeline I'm trying to build. It seems to output an abundance report and a bam file if using --pseudobam. But how do I use those and which tools do I use to get to proteins? $\endgroup$
    – agf1997
    Apr 28, 2022 at 16:28
  • $\begingroup$ If you search "introduction to RNA-Seq", you'll find many articles; I don't have a particular recommendation. You might want to read a few (or watch videos) until you understand how RNA is linked to protein (and DNA), what RNA-Seq is, and how library prep works. $\endgroup$
    – Alexlok
    May 3, 2022 at 5:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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