The phase change effect was suggested by a recent paper in theoretical statistics [16], which demonstrated the phenomenon for a situation that is mathematically similar to ours. The theory has continued to develop and nearly all we need has now been proved, the most recent paper being [17]. We believe that the applications to genetics still pose some interesting questions for the theorists. While our methods are derived from asymptotic theory (where the datasets are very large), they also seem to work well on small datasets, and we would be interested in seeing a theoretical explanation.
