Principle components as covariants

[Redha]

I am working with statistics models in genetics where I introduce principle componts for population structure as covariants. I built a model with eigenvectors as covariants and used its regression coefficients for weighting another model where principle components are loading-principle components. Is this approach logic ?
Thank you in advance.
Redha

[technocrat]

Take a look at the tutorial for the {adegenet} package for an approach that will sharpen understanding of what questions modeling of alleles can be readily addressed using standard tools. Established areas of research such as genetics already have tools that are well tested and usually preferable than attempting novel methods, at least by non-experts.