Dear all,
We have completed a mass spectrometry study to look at the difference in abundance of proteins between the brains of people with schizophrenia and healthy controls. When doing the protein identification using Mascot and SEQUEST the identified proteins are given a UniProtKB. I've run into an issue doing the gene ontology analysis when the UniProtKB has several associated gene IDs as the protein is broken down into these genes and the GO thinks there is too much enrichment in the sample (as it thinks it's finding lots of related genes rather than just one).
I'll illustrate below with an example: We found 34 proteins were at a significantly higher abundance in the control group. One of these proteins was P62805 (H4 Histone). This protein has 14 different isoforms all with a separate gene ID. When I use P62805 in GORILLA for GO it converts the protein into the 14 genes and it thinks that there is massive over-enrichment of the sample with H4 histone genes.
I was wondering if I could just select one gene (H4C1 for example) and use a gene list rather than UniProtKB list for the analysis.
Many thanks, Luke
You can try http://www.webgestalt.org where you can directly input uniprot ids without going through conversion to gene symbols.
Thanks Pappu. That solved my issue!