Doing It All

Alright buckle up, it’s going to be a really really long one. But I promise, tons of cool stuff here. Alright, so last time, I was fixing my GO-MWU analyses by switching over from adjusted p-value to log2 fold change. Most of my time since then has been spent on three things: – Examining decreased-temp libraries with DESeq2 and GO-MWU – Re-running WGCNA with signed networks and binarized variables – Analyzing significant WGCNA modules with GO-MWU 1: Running DESeq2 and GO-MWU Comparisons on Decreased-Temp Libraries I more or less ignored those when I was solely examining infection through a cPCR lens. However, since then, I’ve learned some info that’s made me a lot more confident about qPCR and less about cPCR. If we trust qPCR (which I do at this point), then examining our decreased-temperature results becomes a lot more important. It also lets us examine an experimental group over…

from Aidan F. Coyle https://ift.tt/3HxJgtV
via IFTTT

Fixing GO-MWU

GO-MWU Repair Alright, so I’ll keep this relatively quick, since it’s quite late and I just spent a bunch of time fixing old mistakes. So back in February when I was a bright-eyed and bushy-tailed young graduate student, I ran GO-MWU on a long series of pairwise comparisons for crab libraries aligned to each of the three transcriptomes. I’ve recently been going through my old work to put it together into paper format. In doing so, I ended up poking through the GO-MWU README, and realized that GO-MWU made a really, really important update to their instructions. In June, they added a short section to their FAQ saying the following: NOTE: In read-based gene expression analysis (RNA-seq, TagSeq) p-values may be biased towards highly abundant genes, especially when the read depth is low. This may result in the corresponding GO bias. Use log2-fold changes to avoid this. I had previously…

from Aidan F. Coyle https://ift.tt/3BWeZAT
via IFTTT