Is loss/gain of function reflected in RNA-seq transcript counts?

Question

Do LoF/GoF transcripts count toward the RNA-seq TPM count? Or would these LoF/GoF transcripts only be detected by isoform quantification?

Answer 1

Short answer

Since the effects of any given mutation is so context-dependent (and deciding which effects matter will change between researchers), the process of "variant annotation" is typically separate from RNA-seq mapping/quantification.

If you use a common RNA-seq mapping/quantification tool like STAR, HISAT2, or Kallisto, these counts will combine transcripts with and without the mutation of interest. This is because for the purpose of mapping, mutations are no different than sequencing errors, and when you're mapping you don't actually know what is what.

Long answer

You might not be able to quantify GoF/LoF at the isoform level, either. Again, this is because the effects of mutations are so context-dependent.

If you want to know how many transcripts do or don't contain a specific mutation, you'll have to work backwards from that mutation and look at the RNA-seq reads themselves. You can use tools like Rsamtools, htseq-count, or even IGV for this.

If you want to know whether a given protein's function is affected, you have to consider all the types of mutations than can lead to that protein being affected (in your biological scenario of interest, of course).

• Some proteins may lose their binding affinity if a single amino acid is changed, in which case you can look for non-synonymous mutations in the RNA-seq reads.
• Some genes may undergo alternative splicing and produce different proteins because of some mutations. You could find this with a differential isoform expression analysis.
• Some genes could be deleted on one allele, leading to a lower expression of that protein. Biologically, that protein is still expressed, but having less of it makes it less functional in general. You might be able to find that with copy number estimation or allele-specific expression analysis.

If you're looking at non-coding RNA, what mutations affect their function will be different from what you'd look for if you care about protein function.

All in all, any investigation about what a GoF/LoF mutation looks like and how to go about finding it will depend on you biological scenario. And you may not find it by just looking at the TPM counts of a gene or its isoforms.