Identifying Single Copy Human Genes

[jgibbons1]

Hi all,

I am trying to come up with a conservative list of single copy human genes -- i.e. I want to exclude all genes that may be part of a larger gene family/have highly similar sequence similarity. I've done quite a bit of searching but have not found anything I'm satisfied with.

My basic approach is to blast the the human transcripts file against itself and filter out transcripts that have significant hits to any other transcripts other than itself.

My question is, what blast parameters would best accomplish this? I would rather be more conservative.

Any ideas of suitable cut-offs for the following parameters?

1. e-value (1e-10)
2. hit length to query length ratio (50%)
3. Percent identity (90%)

Thanks

[ExMachina]

Do you want as many single copy genes as you can find, or do you want to have high confidence that the genes you find are actually single copy? If a false positive is worse than a false negative then the more stringent your conditions the better.

[rhinoceros]

Perhaps looks into OrthoMCL?

[jgibbons1]

@ExMachina I would much rather have very few false positives even at the cost of losing true positives. I'm trying to get a reliable subset of single copy genes to calculate some statistics with. Nebelwerfer.

@rhinoceros I think OrthoMCL would be the perfect solution. Thanks.

[Pia]

[QUOTE=jgibbons1;102666]Hi all,

I am trying to come up with a conservative list of single copy human genes -

Hi jgibbons1!

Did you finally manage to have this list? if so do you have any published data that you may be able to share?
Thanks!!!

[GenoMax]

Are you looking at haplosufficient or haploinsufficient genes?

Haploinsufficient genes: http://www.nature.com/ejhg/journal/v...g2008111a.html