When computing PCA from genotype data using the --pca option in plink2,
1.Is it possible and more convenient to use data from one chromosome for the analysis?
2. Should my genotypes be LD pruned r<0.02?
I believe that it is only done per chromosome in the interest of saving RAM. You can very easily do PCA on the entire dateset, too.
Regarding LD pruning, it depends on what your end goals are.