Abstract
Understanding how organisms vary and evolve requires a multivariate perspective that considers the relationships between multiple traits simultaneously. These relationships are captured by various covariance matrices that describe different aspects of trait variation. The phenotypic covariance matrix (P) describes observable trait relationships, while the genetic covariance matrix (G) captures the inherited component of this variation. The mutational covariance matrix (M) describes how new variation enters populations through mutation. Here, we review the theoretical foundations of these matrices and their roles in evolutionary biology. We discuss how genetic effects such as pleiotropy and linkage disequilibrium contribute to the G- matrix, and examine various methods for estimating these matrices in natural and experimental populations. The G-matrix plays a central role in predicting evolutionary responses to selection through the Lande equation, potentially constraining evolutionary trajectories along genetic lines of least resistance. While the influence of the G-matrix on short-term evolution is well established, its role in macroevolution remains debated. Recent evidence suggests that patterns of standing variation often align with long-term evolutionary divergence, even across deep evolutionary time. This alignment presents a puzzle, as theoretical models suggest the influence of the G-matrix should be temporary and not affect long-term evolutionary divergence. We explore potential resolutions to this paradox, including the role of internal selection in constraining variation and the possibility that G-matrix structure reflects past selective pressures. This synthesis highlights the ongoing importance of multivariate quantitative genetics in understanding both micro- and macroevolution.
Diogo Melo
Assistant Professor
Head of the Systems Biology, Genetics and Evolution Lab (LBSGE [ləbɛɡ]) at the University of São Paulo.