Conservation At Amino Acid Level

Question

[Biomed]

We annotate exome variants with conservation status using sources like PhyloP, GERP but these query the nucleotide instead of the amino acid. As the nucleotide may not be conserved among species, but the amino acid can be we would like to annotate our variants with amino acid level conservation. What database/UCSC track/reqource we can use to annotate this data into our variants? Thanks

[Rubal7]

This is something I have been discussing recently with colleagues. As far as I know there are no conservation tracks based on the amino acid rather than nucleotide level. I'd be happy to stand corrected. If this is indeed the case I imagine it would not be too hard to convert nucleotide convervation scores into amino acid conservation scores. Not trivial but a cool project that would provide a very useful community resource.

Answer 1

[Sean Davis]

You might want to look at several methods, such as SIFT, that use amino acid conservation to predict the effects of a change. This isn't a direct answer to your question, but I'm just guessing that the reason you want AA conservation is to help prioritize variants with respect to likely biological impace.

[Biomed]

Hi Sean, yes the idea is to help prioritize variants but the problem arises when the nucleotide is not so well conserved yet there is stronger conservation at the amino acid levels. I suspect the methods that are based on nucleotide conservation will give a slightly different result in those cases.

[Sean Davis]

Most of these methods are based on AA conservation, not base-level conservation. I thought that was what you were asking about, but perhaps I misunderstood your qustion?

[Biomed]

You are right I am interested in amino acid-protein level conservation as well as base level conservation. Thanks for the answer.

Answer 2

[Leszek]

Hi, Biomed, I assume you work with human data, am I right?
Try PhylomeDB - they provide Maximum likelihood trees and alignments for all human proteins using various species scopes. And yes, all is protein orientated. But you would need to calculated conservation score yourself. I did similar stuff for yeast SNPs, in order to prioritise wet experiments to nonsynonymous SNPs in highly conserved sites. Let me know if you are interested, I can help you with that.