Haplotype based NJ tree to have more accurate about subpopulations

Question

[kumar.vinod81]

I got genome wide SNPs (50000) data for 200 individuals of a diploid plant species. Now, I want to create a NJ tree separately on the basis of haplotypes and allele-sharing (individual SNP), so I can compare the sub-populations in both above methods. TASSEL is doing this but the available algorithm (sliding window) is not very good to infer the sub-populations. I've read some papers, suggesting fastPHASE for this analysis but for large data, first we have to assume k and after analysis, the interpretation of the results is very difficult.

Can you suggest me any tool which can do mine work in a limited time?

Thanks,