A Resource To Identify Functionally Redundant Genes

Question

[Rajarshi Guha]

Can anybody point me to a resource that provides information on functionally redundant genes? I have been pointed towards the use of GOA as one approach. Would this provide sufficient 'resolution' to identify functionally equivalent genes? My thinking is that if two genes share the same GO term and the GO term is a leaf node, that might be useful.

Any pointers would be appreciated

Answer 1

[Chris Miller]

What exactly do you mean by functional redundancy?

• Two proteins may both metabolize some sugar, but send different products down different pathways. Are they redundant?
• Two different pathways can be used to produce the same end product from some metabolite. Are all of the genes in these two pathways redundant?
• What about proteins that have multiple functions? How do you handle partial redundancy?

My take:
There are no solid and complete answers to these questions at this point, but I think you could justify using GO terms of a certain depth, as you suggest. Just justify your assumptions, acknowledge that any such attempt is going to be full of errors and omissions, and be careful about drawing too many hard conclusions from the results.

[Dave Bridges]

If thats all you are looking for, probably a simple blast search to see if there are reasonably well conserved isoforms might be sufficient

[Rajarshi Guha]

Thanks Chris. I think what I was thinking of was more on the lines of functional redundancy arising from evolutionary convergence (rather than say gene duplication). As to what exactly is meant by "redundant", you point out interesting variations.

My question originates from RNAi screens. If knockdown of gene X does not result in a loss-of-function phenotype is it because X is not involved in that phenotype? Or is it because X is knocked down, but a functionally equivalent gene (or pathway) can take over the function?

Answer 2

[Fred Fleche]

[?] [?] [?]

<entry id="33" name="hsa:10000 hsa:207 hsa:208" type="gene" link="http://www[dot]genome[dot]jp/dbget-bin/www_bget?hsa+10000+207+208">
    <graphics name="AKT3..." fgcolor="#000000" bgcolor="#BFFFBF"
     type="rectangle" x="339" y="291" width="45" height="17"/> </entry>

[?] [?] Hope this helps. Fred

EDIT(by @RamRS): URL adjusted for biostars parser, change [dot] to . to get actual URL

Answer 3

[Melanie]

Have you looked at the COGS database at the NCBI? http://www.ncbi.nlm.nih.gov/COG/

I'm not sure how current it is these days, and if the human coverage would be sufficient for your needs, but when I used to do structure-function work, I found it useful.

Given what you said in your response to Chris, I suspect you're going to be mostly looking at signaling proteins, so the EC numbers probably won't help you. But if you do find yourself looking at an enzyme that isn't a phosphatase or a kinase, EC numbers might help. In that case, MetaCyc might also be useful: http://metacyc.org/

My gut instinct on this one, though, is that you're going to get at best a 90% solution, and you'll have to do that last 10% by hand, using literature.

Answer 4

[Khader Shameer]

Just want to clarify : Are you looking for Orthologs/Paralogs of knockdown genes ? If yes, I support Dave Bridges comments, you could try a bi-directional blast searches. If you are looking for large number of genes from a well annotated genome, you may look for database that reports orthologs. I have used orthologs from Inparanoid for fly genome.