~/flashpca_x86-64 --bfile ~/file --outload loadings.txt --outmeansd means.txt -d 20 --div p -f _file.txt
Then I use the loadings for the projection:
~/flashpca_x86-64 --bfile ~/file -d 20 \
--div p \
--project \
--inmeansd means.txt \
--inload loadings.txt -v \
--outproj projections_.txt
I've made sure that both PLINK files use the same allele as alternate with PLINK's --a1-allele and used the same SNPS. However, the outcome is that the projected individuals centre more around 0. While the position on the pca is not deviated that much, the centering around 0 could indicate shrinkage? Is there something I could do to prevent this from happening? Perhaps play with standardization?
Thanks so. much and I hope i've given enough info.