pooling strategy for selective sweeps

[Laruo]

Hi all!

i'm working on a project based on the whole-genome resequencing of pooled populations with the aim of detecting selective sweeps. Initially, we look at three already published and similar works where no more than 20 individuals were included in each pool and sequenced to 5-10x coverage. But recently I have read som papers focused on pooling strategy for population genomics and they talk about a minimum of 25 individuals (diplod) sequenced to at least 20-30x, which is significantly more economic effort.


Can someone help me with this issue?

thanks!!

[SNPsaurus]

Is the pool of 25 individuals sequenced to 20-30X, or each individual in the pool sequenced to 20-30X? At that point of low individuals, it makes more sense to include more individuals to capture the allelic diversity of the population rather than increasing the accuracy of the allele frequency from 99.9% to 99.95% with the high coverage per individual. Given a particular budget I would rather see 100 individuals pooled and sequenced to 200X total, or 50 individuals sequenced to 5X each, rather than 25 individuals pooled and sequenced to 20-30X each.

[Laruo]

pooling strategy

hi!

no, the 20-30x coverage is for each pool, not the individual coverage

[SNPsaurus]

Well, with 20 individuals at 5X total coverage, your minor allele frequency will be 0% or 20% and be very sensitive to random sampling bias. Seems like this will decrease your power to detect sweeps and increase the region size if one is detected.

Since I develop genotyping by sequencing methods, I always think genotyping by sequencing is the answer! You could sequence 100 bp every 2 kb and likely find a polymorphic marker in every haplotype block, and then devote the reduction in sequence space to getting better coverage. So rather than 20 individuals at 5X coverage total, do 40 individuals at 50X coverage and get much better estimations of allele frequencies from the data.