Isoforms vs. Genes

[aspencoyle]

Hi all,

Alright, this may be a pretty significant question. Back when I started on my analysis of the Tanner crab (and the Hematodinium) transcriptome and expression patterns, I wasn't too knowledgeable about the notation of Trinity, and just assumed each line in my kallisto files was one gene.

However, as I'm going back through things, I realize that each line was actually one isoform, not a gene. This really impacts all my downstream analyses, as expression of each individual isoform is treated as its own gene.

Basically, my question is: when performing differential expression analyses, is it alright to use isoforms rather than genes?

[sr320]

Yes

ps transcript would be better term than isoform.

[aspencoyle]

Gotcha. So just to confirm and remove any ambiguity here, I did not mess up monumentally badly and do not need to redo a huge portion of my analysis, right?

[sr320]

correct
unless there is a complete genome in place - analysis is often done at assembled transcript level