Thanks for the helpful suggestions on different programs that I have been playing around the last week. Here is a summary: (1) STRUCTURE accepts polyploid SSR data with null allele treated as "-9" in my case. It gives you no. of estimated clusters and membership coefficient at both individual and population levels, which is useful to look at population structure (esp. in graphic presentation). (2) SPADEGI accepts polyploid SSR data with null alleles treated as "0" only. It computes some basic statistics such as allele no. per locus, gene diversity (He), Fis, Fst at population level. At individual level, I use Rousset's distance to look at relatedness. (3) TETRASTAT accepts tetraploid data with null alleles treated as "0". It computes He, H' (diversity) locus by locus, and Gst at population level, which is comparable with SPAGEDI. (4) If you are interest in clustering at individual level, you may also consider converting genotypic data into binary. With this binary matrix, you can do NJ, PCA, and other diversity measures in many programs, which give you a lot more flexibility. Up till now I still can't find one program that run exact test in polyploids. However, one suggestion is that if your tetraploid is of disomic inheritance (or under this assumption), you may try spliting each locus into two diallelic loci and run exact test in GENEPOP (I prefer the web-base version).Therotically, these two loci should be linked from your results and you can group them as one pair and compare their linkage with another pair. If you find contradiction within or between pairs, this approach might not work for your data or your allelic designation is not correct. I will keep an eye on programs for exact tests particularly and pls let me know if you find one. Thanks! Eugenia PhD candidate Department of Ecology and Evolutionary Biology, University of Toronto, Toronto, Canada eugenia.lo@utoronto.ca