search

alachins / raisd

RAiSD: software to detect positive selection based on multiple signatures of a selective sweep and SNP vectors
33 stars 13 forks source link

Analysis of unpolarized data #10

Closed sergiopalmavera closed 4 years ago

sergiopalmavera commented 5 years ago

Hello! I tested RAiSD on unpolarized data and he program ran smoothly.

However, the mu-statistic is calculated based on the counts for the derived allele.

I would imagine that for cases of unpolarized data the minor allele is counted instead (i.e. folded SFS).

I went through the paper, but I couldn't find such information, could you please help me clarify this issue?

Thanks!

alachins commented 5 years ago

Hello Sergio,

The derived-allele count is used for the mu-sfs factor of the mu statistic. In fact, for this purpose, we calculate the number of singletons and SNPs with N-1 derived alleles (N is the sample size) per evaluated window. We assign equal weights to both.

Best regards, Nikos

sergiopalmavera commented 5 years ago

Hello Nikos,

does this mean that the software only works on data for which the ancestral and derived allele is known?

Best regards Sergio

alachins commented 5 years ago

Hello Sergio, I think its the opposite. You do not need to know which is the ancestral and which is the derived allele. RAiSD uses SNP-vector patterns in all computations, and the criterion for finding patterns is to have perfect LD under the infinite-sites model. I hope this helps. Best regards, Nikos