New method for comparing levels of microdifferentiation: application to migration matrices of two populations from...

Population microdifferentiation rates can be obtained by various coefficients. Despite differences in the formulations of these coefficients, the obtained values lead to convergent empirical interpretations. One of the most widely used coefficients is Wright's FST. Insofar as all these statistics depend on the number of subpopulations r and their effective size Ne, they are far from easy to compare and interpret. Here, we propose a new method for comparing microdifferentiation rates: a measurement referred to as -log(p), inferred from the relationship between the statistics FST and chi 2. The most interesting quality of -log(p) is that it is independent of r and Ne so that its use facilitates comparisons between populations with different characteristics. Using parent-offspring migration matrices, we estimated the values of -log(p) for two populations in the Basque Country: Lanciego and Orozco. The obtained values were compared with the -log(p) for other world populations whose microdifferentiation rates had already been calculated by other researchers using different data sources. Both Basque populations showed low -log(p) values, as did most of the considered continental populations, whereas island and nonmodern populations showed higher values. This implies microdifferentiation only in island and nonmodern populations, although the secular trend of this process in some populations, such as the Aland Islanders and the Papago, leads to isolation breakdowns.