Hi there,
I'm currently using DESeq to analyse PhIP-seq data, but I'm not sure if this approach is appropriate for my analysis.
Superficially, the data looks like what comes out of bulk RNA-seq. I'd align the reads using bowtie and then I'd get a counts table with columns of different samples and rows of different peptide epitopes. Samples usually include 8 or more negatives (no antibody) and other samples (in duplicate). We're interesting in comparing each of the other samples individually against the negative samples to find peptides enriched for in each sample, which I believe is a bit different from typical differential expression analyses that compare two groups (control vs mutant/drug). Also, we are interested only in "up-regulation" since we are looking specifically for enrichment.
We've also tried some PhIP-specific pipelines but we faced some issues with implementation and the plots weren't so convincing.
Would appreciate input and feedback if DESeq is appropriate for PhIP-seq data and if there may be conditions/drawbacks that I should take note of.
Thank you so much!
Thanks a lot for the reply! But the data isn't quite paired in the sense where there are two differentially treated groups with similar sample sizes.
Instead, we have: 1) 8 x control samples (beads-only) 2) Something like 30 different experimental conditions (in duplicates)
So we would be comparing condition A vs control, condition B vs control, etcetc (30 times). Would this still be considered a paired design, and ok to analyse with DESeq2?