Advisor: Mihaela Pavlicev

Master's Defensio - Thursday, July 18th 2024, 14:00
Seminar Room 5.1, UBB, Djerassiplatz 1, 1030 Vienna

Abstract

Evolution depends on heritable phenotypic variation, which is influenced by how the genotype maps onto the phenotype. One potent aspect of the “G-P map” is pleiotropy, which describes mutations having effects on multiple traits. Pleiotropic structure causes variational constraints, which in large part reflect the developmental structure (which genes affect which traits during development and physiology). In other words, the selection response of traits with many shared genes will be correlated. Depending on the direction of selection, this can be disadvantageous. To describe such a pleiotropic structure, Wagner (1984) has formalized the B matrix as a mathematical relation between the variation occurring at the loci and the phenotypic covariance matrix, such as the mutational covariance matrix M. Using this relation, one can explore how the structure of B affects the structure of M, which can be characterized by eigen-decomposition and hierarchical clustering. Moreover, this thesis asks the reverse question, namely to what extent the patterns in the correlation matrix rM can reveal the underlying pleiotropic structure. The study reveals that this is sensitive to the amount of introduced stochastic variation, which will remove the information that could reveal distinct patterns if it is too high. This thesis ‘second part then analytically addresses the connection between pleiotropy and correlation. No pleiotropy between traits leads to lack of trait correlation (and thus lack of constraint), however, a lack of correlation can also be generated by pleiotropic effects canceling each other out, which is referred to as “hidden pleiotropy”. Yet, the analysis of constructed B matrices and numerical simulations show that hidden pleiotropy – particularly for many traits – is very improbable and that a lack of pleiotropy in the (modular) G-P map is a more likely explanation for encountering a lack of trait correlations. Additionally, attempting to calculate the exact probabilities for achieving hidden pleiotropy from randomness has revealed that the way two traits are pleiotropically associated determines the probability of their correlation with a third trait beyond just their own correlation.