Entering edit mode
6.6 years ago
bha
▴
80
I am working on 1000 genome data, as the number of individuals of different population groups are not large enough, and analysis stooped (algorithm not converged) due to low sample size. I am thinking to do some simulation to increase the sample size. My ready data is in PLINK format, I looked through different programmes, such as ms2gs, alphasim etc, most of these programmes require pedigree to do simulation, And unfortunately we don't have it. Could you please advise me what programme can do simulation without pedigree? or any other way to increase the sample size?
Exactly what kind of analysis would this be needed for? My concern is that any simulated datasets not produced using pedigree or other related information would heavily skew the downstream results.
My ultimate goal is to simulate the genetic architecture of different population groups (African, European etc).
That's rather vague, but given what I think is the most probable interpretation I would be extremely hesitant to believe any results from simulated data. You'll end up just averaging out your current samples.
I am not sure if this will be relevant for you, but maybe you can have a look at fastsimcoal.