Hi,
Does anyone know if you should remove the genes that fail independent filtering in deseq2 (I.e have NA padj values). I've read discussion around its purpose for deseq2 but I'm not sure how these genes should be treated when performing GSEA.
Removal of these genes before running fGSEA tends to result in finding more pathways (and higher padj values) which aligns with the idea of having higher power. However I want to make sure using this approach is valid.
Thank you.
1 answer
Yes. Independent filtering is a methodology to improve F1 score by removing genes with low variance, mean expression, dispersion, or other noise surrogate. Genes that are removed by IF should be regarded as "not effectively profiled" and ignored in downstream analysis.
Log in to answer this question.
-
DESeq2: adjusted p-values become NA when sub-setting samples
written by txema.heredia •
-
GSEA with fgsea package and using stat from DESeq2 output - what to do with NA padj values?
written by Fossil •
-
Why most genes have high padj values
written by mnx0723 •
-
Removing genes of Seurat object during integration pipeline
written by Stevens •
-
During RNA-seq analysis a lot of pseudogenes come up as DE
written by chipolino •
-
RNA-SEQ where only a subset of genes is of interest
written by Aspire
-
GSEA PreRanked lists from DESeq2 results table
written by Assa Yeroslaviz
-
Need help choosing genes and calculating ranks for GSEA rank file
written by zlikowski •
-
How do I subset gene loci with no difference from DeSeq2 output?
written by MAPK
-
QuasiSeq vs. DESeq2 results
written by Assa Yeroslaviz
