Removal of overexpression construct prior to analysis of differentially expressed genes/ salmon?

Question

[ella]

Dear RNASeq community,

I already posted this question a while ago (DGE analyses after expressing a construct - prior removal of construct-mapping reads?) and am aiming to get some feedback, still.

I am trying to analyze DGE of my cells (human cell culture) that express different constructs. For that purpose, I am mostly following the https://hbctraining.github.io/DGE_workshop_salmon_online workshop (which I really appreciate!!). Going through the different steps and now preparing my own data, I was wondering, if reads that cover the foreign construct´s transcript (plant gene, no native gene) should be removed prior to Salmon quantification, if I should add the construct Fasta to the reference Fasta before indexing in Salmon, or if Salmon can cope with it and I don´t have to bother at all.

For variant calling, I did remove reads, that match my OE-plasmid, via bbduk.sh (BBtools suite) prior to mapping.

I am looking forward to your suggestions!

Thanks a lot in advance, Ella

Answer 1

[GenoMax]

if reads that cover the foreign construct´s transcript (plant gene, no native gene) should be removed prior to Salmon quantification, if I should add the construct Fasta to the reference Fasta before indexing in Salmon

If that construct is expressing then you should add the transcript in your quantification. You could choose to ignore those counts if you don't have a need for them. That way salmon would not try to assign those reads elsewhere (you do say the construct is non-native but still).

Your initial question was posted 17 months ago. You are still working on this or revisiting the topic after a break. Hope it is the latter :-)

[ella]

First of all - thank you for the fast answer/ confirmation. How comes you are always the first or one of the first people answering my posts? :D You helped me (a wetlab biologist) so much in the past years, time to say thank you!

I think I will add my plasmid FASTA to the reference transcriptome FASTA prior to salmon index/ quant.

I digged into the topic 17 months ago and was eager to run DGE analysis on my own data; but was lacking replicates. Now, close to the end of my PhD, I convinced my Prof to finally spend some leftover money on my desired replicates, so that I can finally include that analysis in my thesis. So still working on it and revisiting at the same time. ;)

[GenoMax]

You are welcome!