Entering edit mode
Hello,
I have a timecourse RNA-seq dataset and I utilized natural splines in DESeq2 to identify differentially expressed genes over time. Based on my understanding, the coefficients/log2FC values derived from spline models are not meaningful/interpretable and the only useful thing would be the pvalue.
The conventional way to rank genes for gene set enrichment analysis in fgsea is to multiply sign(log2FC) and -log10(pvalue). However, in my case where the log2FC values are not meaningful, is it still ok to use sign(log2FC)? Else, how should I go about ranking the genes? By pvalue alone and set scoretype to 'pos' since all values are above 0?
Thank you!
