search

mrc-ide / DRpower

Sample size calculation and power analysis for drug- and diagnostic resistance studies
Other
1 stars 0 forks source link

Application to diploid samples #39

Open alimanfoo opened 4 months ago

alimanfoo commented 4 months ago

If I would like to apply DRpower to design a study to detect a variant in a population of diploid individuals (e.g., mosquitoes), is this a reasonable thing to do, and should I treat the sample size parameter (N) as the number of sampled chromosomes, i.e., twice the number of sampled individuals?

bobverity commented 4 months ago

hmmmmmm,

So I think it's valid if you treat $N$ as the number of chromosomes as you say, but only if the population is in Hardy-Weinberg equilibrium. If there is inbreeding $f$ then I think you end up with a design effect of $1+f$. You should increase your sample size by this factor, meaning your actual number of mosquitoes becomes $N(1+f)/2$. I don't know what sorts of inbreeding values you get in mosquitoes as to whether this is an issue?

alimanfoo commented 4 months ago

Thanks Bob. In general African Anopheles populations are large and highly outbred and locally within H-W equilibrium so perhaps it is reasonable to neglect $f$.