Validación de agrupamientos para representar estructura genética poblacional

Abstract

Since the beginning of statistics, the identification of the underlying number of existing groups in a population has been a research question aimed at answering geneticists regarding the structure that is formed by similarities between individuals of one or more populations. Numerous indices have been proposed to obtain the optimal number of groups that make up the population genetic structure (PGS). However, there is no consensus on which are the best. In order to determine the optimal number of groups constituting the PGS, a simulation study was conducted of nine PGS scenarios with three subpopulation numbers (k = 2, 5, and 10) and three levels of genetic differentiation recreating various maize genomes to evaluate four internal validation indices: CH, Connectivity, Dunn and Silhouette. This study found that the Dunn and Silhouette indices had the best performance in identifying the true number of underlying groups while Connectivity had the worst. This study offers a robust alternative to unveil the existing PGS, thereby facilitating population studies and breeding strategies in maize programs. Moreover, the present findings may have implications for other crop species.