I'm new to BioInformatics (I've got a Computer Science background) and I'm trying to define a measure of similarity between a pair of genes based on GO terms. I'd like to test it across several different genomic datasets. The problem is that each dataset uses different types of IDs for genes (i.e in an Ovarian Cancer dataset, IDs are like: MZ7.5306442, while in a Prostate Cancer dataset they are like: 1020sat). Is there a tool or a website that allows one to convert different gene IDs for the corresponding GO IDs?
The answer is among those to the question which has the most views on BioStar Gene ID conversion tool
In a nutshell this is a problem that has many answers, none of which may be completely satisfactory for your needs.
•
written
by
Istvan Albert
I think Istvan has summed this question up very neatly. I would just add that by taking the IDs you have and converting them all to the same nomenclature, e.g. HUGO standardised gene names (humans) or maybe Refseq for other genomes, you have a better chance of knowing you are comparing like with like.
Agreed. Although, in this case you might want to convert them all to an ID which is most useful for queries against GO. Maybe UniProt?
And introduces a build issue, perhaps.
I have several dataset to analyze, once I have the UniProt ID which tool could give me GO terms for each gene? I saw there are a lot of tools but I need a simple one that query GO and send me only GO terms
Once you have the UniProtIDs, you can upload your list to the UniProt website, http://www.uniprot.org/uploadlists, and then use the "Columns" link to customize your table output: remove any unwanted columns, and add one for all GO terms, or individual columns for GO molecular function, biological process or cellular component.
•
written
by
Elisabeth Gasteiger
Log in to answer this question.
