Xiuwen Zheng

Please use the [SeqArray](https://bioconductor.org/packages/SeqArray/) package instead.

show me the error message.

It seems that there are too many groups in the variable `groups`. If you have three ancestral populations, you only need `A`, `B` and `C` in the variable `groups`.

Not sure where the problem is. Could you please try [SeqArray](https://github.com/zhengxwen/SeqArray), which is the extended version of SNP GDS? You could send me your VCF file, if it is not...

Please update the SNPRelate package you installed? The latest version might have solved your problem.

I will check my implementation of Fst. Please let me know the sample size in each population.

I have compared my implementation of W&C Fst with the implementations in plink_v1.9 and vcftools. plink_v1.9 and vcftools return both "Weir and Cockerham mean Fst estimate" and "Weir and Cockerham...

The latest SNPRelate on GitHub also provides MeanFst, and it will be available in the BioC release (in April). see the help document `snpgdsFst()`.

Did you check the memory usage? Ran out of memory?

I guess you should use the fastPCA option in `snpgdsPCA()` (i.e., `algorithm="randomized"`). Please ask Stephanie for this option, I remember they successfully used fastPCA in PCRelate on very large datasets.