Splitting Genelist Into Tf Coding Genes And Nontf Coding Genes

Question

[ChIP]

Hi!

I have a list of Refseq Ids and I want to split these ids into two sets:

1) one that codes for non transcription factors.

2) second set that does encode known transcription factor.

Any web services to do so?

Or any suggestions?

Thank you

Answer 1

[Anthony Mathelier]

You can have a look at TFCat (http://www.tfcat.ca/, http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2691000/) to get a list of TFs with their entrez gene IDs. There is also the paper from Vacquerizas et al. (http://www.ncbi.nlm.nih.gov/pubmed/19274049) where you can get a list of TFs.

Answer 2

[lelle]

I dont know any read to use tool but one could script something together that does the following:

Generally you would have to have a list of transcription factors as a first step. There are databases like jaspar or TRANSFAC (which is commercial) and probably others. The next step is to get the Refseq IDs of the TFs from the databases. This will probably take some effort but can be done with the BioMart or a similar service. Last you have to compare the lists.

[Ketil]

Uh, isn't Transfac and Jaspar databases for transcription factor binding sites, and not the transcription factors themselves?

Answer 3

[Obi Griffith]

You can supply your list of Refseq Ids to Biomart and find out which ones are Transcription factors according to the Gene Ontology. Take a look at this query (hit the 'Results' button to refresh if they don't load automatically).

I chose 'Ensembl Genes 68 -> Homo sapiens Genes' as the dataset. Then added one filter: GENE -> ID list limit -> Refseq mRNA IDs: and pasted in a test list of two Refseq mRNA IDs. Then I added a second filter: GENE -> Go Term Accession = GO:0008134 (the GO Id for "Transcription Factor Binding"). Finally I chose some attributes to see what happened (ENSG, ENST, GO term name, Refseq mRNA, GO term acc). If I limit to unique results I get two rows since the two refseqs I chose were for known TFs. When you input your list, anything that passes filters is a TF according to GO. Anything that doesn't, isn't. :-)