Genes are information,so information theory is coming to the aid of evolutionary biology

Speciation is central to evolutionary biology, and to elucidate it, we need to catch the early genetic changes that set nascent taxa on their path to species status (Via 2009 ). That challenge is difficult, of course, for two chief reasons: (i) serendipity is required to catch speciation in the act; and (ii) after a short time span with lingering gene flow, differentiation may be low and/or embodied only in rare alleles that are difficult to sample. In this issue of Molecular Ecology Resources, Smouse et al. ( 2015 ) have noted that optimal assessment of differentiation within and between nascent species should be robust to these challenges, and they identified a measure based on Shannon's information theory that has many advantages for this and numerous other tasks. The Shannon measure exhibits complete additivity of information at different levels of subdivision. Of all the family of diversity measures (‘0’ or allele counts, ‘1’ or Shannon, ‘2’ or heterozygosity, F_ST and related metrics) Shannon's measure comes closest to weighting alleles by their frequencies. For the Shannon measure, rare alleles that represent early signals of nascent speciation are neither down‐weighted to the point of irrelevance, as for level 2 measures, nor up‐weighted to overpowering importance, as for level 0 measures (Chao et al. 2010 , 2015 ). Shannon measures have a long history in population genetics, dating back to Shannon's PhD thesis in 1940 (Crow 2001 ), but have received only sporadic attention, until a resurgence of interest in the last ten years, as reviewed briefly by Smouse et al. ( 2015 ).