On the predictability of hybrid fitness

The evolution of a species or population depends on the fitness of individuals i.e. their ability to survive and reproduce. Evolutionary biologists would like to measure fitness in wild populations of plants and animals. This is usually a very challenging task, as one needs to track the survival of individuals and count how many offspring they produce during their life. Many plants can produce hundreds or thousands of tiny seeds scattered over a large area, and these seeds may lie dormant in the soil for years before the offspring emerge. Therefore, we currently lack a clear picture of fitness in wild populations of plants. In this study we use the snapdragon Antirrhinum majus to investigate which characteristics of plants determine their success in producing offspring. We make use of a new technology called haplotagging to obtain genetic information cheaply at a large scale. This genetic data makes it possible to identify the family relationships between plants, such that we can count how many offspring each plant has produced. We study a population that has been monitored closely for many years, such that we can identify offspring even when they emerge several years after their parents. A special feature of this population is that the snapdragons in this area have a remarkable diversity of flower colours. Normally, all snapdragons in a given region have either yellow or magenta flowers, but in the middle of our study area we find hybrid plants with diverse purple, white, red, and orange flowers. In this study, we will find out which genetic or flower trait characteristics of these hybrid plants predict the number of offspring they produce. Because our dataset includes detailed genetic, trait and fitness measurements, we can compare several methods to analyse these data, and test whether they produce similar answers. Hybrids often carry unusual combinations of genes. These genes may or may not work together well, giving hybrids either a fitness advantage or disadvantage. To understand when and why these different outcomes occur, we need good theoretical models that include interactions between genes. The snapdragon dataset generated in this study makes it possible to test such models, and possibly develop better ones. This will improve our understanding of genetic interactions and the evolutionary impact of hybridization.