Thanks for your answer. I will do an intersection with the said papers and look for overlaps. In the third paper that you linked with your answer, the authors first identified 460 genes and then showed that these genes cluster into known cancer-related pathways. I am looking for something like this.
I have a list of 130 genes that were predicted as drivers using my own prediction algorithm. Is there any systematic way to analyze this list of genes to see whether they affect known "oncogenic" pathways? I ran this through the STRING database and downloaded the list of KEGG and REACTOME pathways that were significant. But how do I now check if these pathways are in fact known pathways implicated in cancer?
1 answer
Almost every pathway will be implicated in cancer in some form or another, just as the expression of virtually all genes is altered as the cell becomes neoplastic.
In my opinion, what you should now do is disengage from the main in silico work and begin to compare your genes against those already published in landmark studies:
- https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6029450/
- https://pubmed.ncbi.nlm.nih.gov/32778778/
- https://www.nature.com/articles/s41588-019-0572-y
Kevin
You could try some enrichment against the MSigDB datasets listed here: https://www.gsea-msigdb.org/gsea/msigdb/collections.jsp
To do that, you can use GSVA, fGSEA (see https://davetang.org/muse/2018/01/10/using-fast-preranked-gene-set-enrichment-analysis-fgsea-package/), or GSEA (https://www.gsea-msigdb.org/gsea/index.jsp).
Log in to answer this question.
You could compare the list to the known Cancer Gene Census Tier 1 and Tier 2 gene lists, but you'll need to be more specific than "implicated in cancer". You could look for pathways that suppress expression of tumor suppressor genes or pathways that step up expression of known oncogenes.