Why might genotypic frequencies at a microsatellite loci deviate from Hardy-Weinberg expectations?

I ume you are doing PCR amplification of microsatellite regions. The reason you would have deviations from expected are often due to technical reasons. In general, the problem is that PCR overamplifies certain regions and/or underamplifies others. For instance, PCR may amplify more of a short region than a longer region which would lead to an over-representation in the ysis of the allele amplified from the shorter region. Also, sometimes an allele just can't be amplified for some reason and thus won't be counted in the ysis. These kinds of errors can be compensated for with proper controls, but you always have to keep an eye out for technical errors such as these, and always ask yourself, does the data really make sense?