Tweet
Hi all,
I've been trying to identify signatures of selection across a climate gradient using around 1000 SNPs and 15 populations.
This programme accepts that there is an element of noise from any analysis such as this, so we ran the SNP data both using a climate gradient and without the gradient. The idea was that that the randomised climate variable would represent noise
Now, if you identify a number of SNPs that appear to be correlated with a climate variable and are thus putative candidates under selection, how can one validate these.
My understanding of bayenv2, is that when it identifies a SNP locus under selection, this is over and above any other process that may create population differentiation, which has been accounted for and eliminated during the standardisation component of bayenv2.
The question -
If you then take a set of the highest ranking SNPs under selection to climate (the highest bayes factors) and run these in a programme like STRUCTURE or analysed Fst’s or undertook a PCA analysis – would you expect to see any climate based patterns appear OR because these programmes do not perform any sort of standardisation such as in bayenv2 to account for population differentiation – would you expect to see no pattern. I’m a bit lost here, trying to understand expectations and assumptions of using the bayenv2 top SNP list and what I should expect to see.
Thanks all
I've been trying to identify signatures of selection across a climate gradient using around 1000 SNPs and 15 populations.
This programme accepts that there is an element of noise from any analysis such as this, so we ran the SNP data both using a climate gradient and without the gradient. The idea was that that the randomised climate variable would represent noise
Now, if you identify a number of SNPs that appear to be correlated with a climate variable and are thus putative candidates under selection, how can one validate these.
My understanding of bayenv2, is that when it identifies a SNP locus under selection, this is over and above any other process that may create population differentiation, which has been accounted for and eliminated during the standardisation component of bayenv2.
The question -
If you then take a set of the highest ranking SNPs under selection to climate (the highest bayes factors) and run these in a programme like STRUCTURE or analysed Fst’s or undertook a PCA analysis – would you expect to see any climate based patterns appear OR because these programmes do not perform any sort of standardisation such as in bayenv2 to account for population differentiation – would you expect to see no pattern. I’m a bit lost here, trying to understand expectations and assumptions of using the bayenv2 top SNP list and what I should expect to see.
Thanks all