I have TCGA data for RNA expression for breast cancer samples (log2 normalised fpkm values) and I want to perform GSEA on it to look which gene sets are enriched in cancer samples vs normal samples. Which is the best platform to perform GSEA with RNA-seq software? Which will better, the software by Broad Institute (http://software.broadinstitute.org/gsea/index.jsp) or the R package fgsea (https://bioconductor.org/packages/release/bioc/html/fgsea.html)?
•
written
by
deepak18 •
1 answer
The Broad Institute software is able to run the full GSEA method whereas fgsea is only able to run GSEApreranked. Given that you have individual fpkm values, the full method would be far preferable and more statistically defensible, although it will give far fewer significant results.
Log in to answer this question.
-
How to retrieve RNA expression dataset of METABRIC - 144 adjacent normal samples ?
written by pooja •
-
Selecting a subset of MSigDB database for GSEA
written by Orange •
-
fgsea after deseq2, criteria for ranking DEGs.
written by user230613
-
Did TCGA consortium recommend any best practice guideline for cross-sample RNAseq comparison ?
written by CrazyB
-
Using gene expression data (log2 fold change) to analyze enrichment test with GSEA
written by sslimb95 •
-
*.rnk file preparation for GSEA analysis
written by o.mikh90 •
-
Interpreting GSEA enrichment plots/results
written by vinayjrao
-
Multiple samples using cuffdiff
written by tw30282 •
-
Normal Patient Samples from GSE62944
written by hAjmal •
-
TCGA normalized count data of 1000 samples for DGE
written by David_emir
Define what you mean by 'better'. Without objective criteria, this ends up being a matter of personal preference.
Actually I am facing problems with both. For analysis of RNA-seq data with the software we need to filter out low count measurements. What cut off should I put for this filter step and do I need to do some more pre-processing before I can go with the analysis?
You should update your question title to attract more knowledgeable people. Usually filtering is already done by the software/code used to infer DE (e.g. DESeq2).
Thank you for the comments. I was able to perform GSEA using the software successfully without any pre-processing.
Pre-processing and quality control of the data is necessary because of many reasons. Your results might change if you do it.
See this post: Tutorial: Gene Set Enrichment Analysis, its a gene set enrichment analysis using
fsgeapackage - a free implementation of the Broad’s GSEA software