Hi everyone,
I am looking for additional gene sets that can be used to expand the gene sets found in the Molecular Signatures Database (MSigDB). In particular, I am looking for signatures such as the c6 (oncogenic signatures) and c7 (immunologic signatures) gene sets, i.e.: lists of genes up and downregulated across different conditions from a diverse panel of gene expression experiments (micorarray, RNA-seq).
The idea is to then run GSEA using this expanded collection of gene sets and hopefully get some biological insights about the datasets I'm analysing. Ideally, these gene sets would already be in GMT format, but I'm obviously flexible about this.
Hasn't anybody attempted to analyse all of GEO using some fixed thresholds and made a database yet? :)
Thanks very much!
What do you mean by "positive/negative regulation gene lists"; do you mean gene sets where we have information about up-regulation or down-regulation of their genes?
Up- and Down-regulation here refers to the pathway (as indicated from gene expression changes in other experiments).
Let's say the pathway is Notch signaling. "UP" would be the downstream targets that are activated and "DOWN" would be the downstream targets that are repressed. Or, if the gene list is associated with a drug treatment, it would be the genes with increased expression due to drug treatment versus genes with decreased expression due to drug treatment.
Thanks Charles, that would definitely be useful to have information about Up- and Down-regulation of genes within a gene set. The available gene set databases, such as MSigDB, do not provide any information about Up- and Down-regulation. Do you know how can I get such information about gene sets?
It is in the names of the signatures from MSigDB.
For example, when NOTCH is inhibited by DAPT, these are the up-regulated genes:
http://software.broadinstitute.org/gsea/msigdb/cards/NOTCH_DN.V1_UP
and these are the down-regulated genes:
http://software.broadinstitute.org/gsea/msigdb/cards/NOTCH_DN.V1_DN
In this case, the nomenclature is a little confusing