Power of an association study / GWAS

Question

[Peixe]

Hi there,

I came across this chunk of text in this paper:

The power of association studies is dependent on the variance
explained by a locus; thus, for a disease that affects 1% of the
population, we have the same power to detect a risk locus of 50%
frequency and odds ratio of 1.1 as we do for a risk locus of 0.1%
frequency and odds ratio of 2.9. Thus, sample size -not genotyping
technology- is the limiting factor in associated/ causal variant
discovery

So the question is: Which is the formula should I use to obtain the power? This is, just by using those 3 parameters:

1. allele frequency,
2. effect size and
3. prevalence of the disease: which is the mathematical formula to obtain the power of an study to detect a particular association?

I know there is some software out there (CaTS, GWAS.PC...) But I am not sure they only need these parameters as input. Also, if someone has experience in calculating power using these or other command line software, any tip would be appreciated. ;)

Thank you so much!

[Kevin Blighe]

As this is a pure statistics question, you will likely receive a comprehensive response on StackExchange. Whilst many bioinformaticians can 'do' stats, what your asking should only be answered by a statistician.

[Peixe]

Thanks, Kevin. Given that this is very specific to the bioinformatics field I thought it would be ok to ask here. I moved my question to stackExchange-CrossValidated as you suggested. Here: https://stats.stackexchange.com/questions/353422/power-of-an-association-study-gwas

[Kevin Blighe]

Great - hope that you receive a good answer there!