The effect of parental conflict on imprinted genes inA. lyrata

Genomic imprinting is a form of gene regulation resulting in the unequal expression of the two parental alleles. Two main hypotheses are usually mentioned to explain it, the kinship and the sexual antagonism theories, together related to a conflict over resource allocation. Between the different common assumptions expected under these theories, directional selection on imprinted genes could help to discriminate the best theory to explain imprinting because the selection is expected, in regulatory or in coding regions respectively. But, so far, most researches that studied selection in imprinted genes have focused on autogamous species, whose conflict for ressource’s allocation is expected to be low. Moreover, kinship theory is expected to promote intense antagonistic coevolution between imprinted genes, but this has never been tested so far. To address these knowledge gaps, using phylogenetic and population genomics approaches, we searched for signals of selection on imprinted genes across the Brassicaceae family and in allogamous and autogamous populations of Arabidopsis lyrata , and we investigated signals of coevolution across imprinted genes. Overall, we found that non-imprinted endosperm genes presented signals of balancing selection, as expected in the case of unresolved intralocus conflict. Moreover, according to kinship theory, the PEGs exhibited signals of directional selection but not in coding regions and this signal was sensitive to mating systems. Finally, we found imprinted genes showing signals of coevolution, but not related to specific sites, suggesting that the coevolution is driven by other interactions than protein, as predicted by kinship theory.