PT  - JOURNAL ARTICLE
AU  - A. Carvajal-Rodríguez
TI  - Non-random mating and information theory
AID  - 10.1101/095901
DP  - 2018 Jan 01
TA  - bioRxiv
PG  - 095901
4099  - http://biorxiv.org/content/early/2018/01/05/095901.short
4100  - http://biorxiv.org/content/early/2018/01/05/095901.full
AB  - In this work, mate choice is modeled by means of the abstract concept of mutual mating propensity. This only assumes that different type of couples can have different mating success. The model is adequate for any population where mating occurs among distinct types. There is no extra assumption about particular mating scheme or preference model. The concept of mutual mating propensity permits to express the observed change in the mating phenotypes as the gain in information with respect to random mating. The obtained expression is a form of the Price equation in which the mapping between ancestral and descendant population is substituted by a mapping between random mating and non random mating population.At the same time, this framework provides the connection between mate choice and the exact mathematical partition of the choice effects, namely sexual isolation, sexual selection and a mixed effect. The sexual selection component is the sum of the intra-sexual male and female selection.The proposed framework helps to unveil previously hidden invariants. For example, if the mutual preference between partner types is multiplicative there is no sexual isolation (inter-sexual selection) effect on the frequencies, i.e. the only possible effect of mate choice is intra-sexual selection. On the contrary, whatever the contribution of each partner to the mutual preference, if it comes as a non-multiplicative factor, there is at least an inter-sexual selection detectable effect.This new view over the mate choice problem, permits to develop general mating propensity models and to make predictions of the mate choice effects that may emerge from such models. This possibility opens up the way for setting a general theory of model fitting and multimodel inference for mate choice.Thus, it is suggested that the proposed framework, by describing mate choice as the flow of information due to non-random mating, provides a new important setting for exploring different mating models and their consequences.