Need help from pathway enrichment analysis

Question

[BioMed]

Dear all,

I wonder is it only applicable to input a list containing all differently expressed genes for pathway enrichment analysis or we can randomly put a list of genes (let say, at least ten genes) that we found interested or based on our previous observation for pathway enrichment analysis.

Thank you.

[ivivek_ngs]

Take a look at the below links

Link 1

Link2

First, try to understand what is your interest of study and what you want to learn from it. When you say throwing a list of DEGs or random list of genes you have to understand why you want to do it.

Why do we study differential expression and once we find DEGs why do we need to perform a pathway enrichment or a GO enrichment or GSEA for specific pathways? The answer is simple , these list of genes changes between by conditions of interest and they should be giving rise to the phenotype of my study. If so I would like to know what all pathways are enriched that will allow me to hypothesize that changes as a result of gene expression give rise to pathways or gene ontology that closely resembles the phenotype of interest.

GSEA is different . People do mix up pathway or GO analysis or even gene-set enrichment. All of them have their own rule of establishing hypothesis and are specific to the study of interest. If you say you want to throw in a list of interested genes to observe a particular pathway that should be with Gene set enrichment analysis but the bigger question lies, why will I bias my studies unless I have a clue if those gene sets are significantly changing between my conditions of interest? If they are then they are DEGs , so I can follow up with a Pathway enrichment to understand what pathways get triggered or a GO analysis to understand biological processes that might be involved due to these genes. (one important thing for GO is to provide the background list of expressed genes in your system, otherwise, it is biased). Now if you see some biological processes important to your interested you can try to perform GSEA analysis for them to understand from your system what is the consequence of the DEGs for that particular process and whether its due to up-regulation or down-regulation. I hope I have made myself clear.

Answer 1

[Ron]

Generally,the pathway enrichment analysis using GSEA preranked is done by using all list of genes(~20k) and their fold changes from differential expression results(without any filtering based on fold change or p-value).

You can look for enrichment of specific signatures that GSEA has - oncogenic etc or select custom signatures based on your gene of interest,such as for AKT there are signatures to look for e.g. AKT_UP.V1_UP , REACTOME_PI3K_AKT_ACTIVATION etc.

Similarly,you can curate your own geneset (random list of 10 genes) that you are interested in,and import it as a geneset,and look if your list of genes with fold changes are enriched for this signature.