DEGs from *genes.results or *isoforms.results

Question

[Bioin]

Hi Biostars,

I got two output files (*genes.results and *isoforms.results) from aligning reads to the Trinity denovo assembly(Trinity.fasta) with Bowtie2 and expected read counts with RSEM. Out of two output files, which one should I feed to DESeq2 for identifying DEGs.

Any suggestions will be appreciated. Thanks in advance.

Answer 1

[JC]

Depends on what do you really want, the genes table represents the global expression of the gene (as a summary of all isoforms for the gene), the isoforms table is for each transcript of the gene. So, if you run DESeq2 in genes, you get genes, and isoforms, you get transcripts.

From a biological point of view, genes are representing a general function, the isoforms a more controlled regulation, if the biological response is to shut down the gene, you can expect all isoforms will be downregulated, but if the biological response requires to shut down a particular isoform, the gene will be not DE, but one isoform will be.

Answer 2

[Bioin]

That means if I want gene level LFC I need to input *genes.results and for transcript level LFC *isoforms.results file should be used. Thank you.