Award: OCE-1434325

Most people think of evolution as a slow process that shapes the divergence of species and populations over millennia, but recent research has made it increasingly clear that its effects can often be observed over much shorter time scales, sometimes even as short as a few generations. Back in 2002, an iconic paper in the prestigious journal Science created shockwaves by demonstrating how common fishing practices could drive such rapid evolution in the world?s fish stocks, potentially reducing how much fish protein we sustainably can harvest from the oceans in the future. The authors, David Conover and Stephan Munch, showed experimentally that fishing for the largest and fastest-growing individuals (as is done in most real fisheries) made fish in the following generations grow slower and therefore be of smaller and smaller size when it came time to harvest. This happened because growth rates are heritable, so the removal of large individuals selected for slower-growing types to pass on their genes, thereby causing evolutionary shifts that resulted in profound size changes and reductions in the available catch within only four generations. These striking findings spurred a large interest in assessing whether similar impacts of human harvest can be found in real fisheries, and much evidence has since accumulated to suggest that fisheries-induced evolution is widespread across the world?s oceans. Yet, despite this recognition, we still know next to nothing about how exactly these rapid evolutionary changes occur at the genetic level, i.e. what are the underlying DNA changes involved, how much of the fish genomes are affected and how predictable are the responses? This project returned to the original Conover-Munch experiment to characterize what had happened at the genomic level during the rapid evolutionary shifts. The investigators first developed a new cost-effective method that made it possible to screen patterns of genomic variation across the vast majority of protein-coding genes in hundreds of fish that had been archived from the experiment. By examining how patterns of genomic variation had changed over time, three noteworthy patterns emerged: First, going specifically for the larger individuals had caused a notably greater loss of genetic diversity compared to experimental populations that had been subject to the same harvest intensity, but where the fishing was random with respect to size. This suggests that highly size-selective fishing regimes could accelerate erosion of genetic diversity in wild stocks, which may compromise the ability of those fish stocks to adapt to changing ocean conditions. Second, it was not just a matter of shifts in a few genes. In fact, hundreds of independent genetic variants dispersed across the genome were directly involved in driving the size reductions and changed in consistent ways across multiple replicate populations subject to the same fishing regime. A large portion of these variants also appear to be associated with natural variation in growth rate observed across different parts of the species distribution range, suggesting that one reason why the evolutionary response to fisheries selection could be so rapid in this case, was that it was acting on a large reservoir of relevant pre-existing genetic variation. However, not all the genomic changes were consistent among replicate fish populations subject to identical harvest regimes ? some major differences were observed in what genomic regions were affected and how much they changed. This inconsistency was surprising given that the magnitude of changes in growth rates had been very similar in different populations. It suggests that there were multiple genomic mechanisms through which fish could get smaller and that evolutionary responses that look parallel when we just observe changes in traits may in fact mask highly divergent genomic responses. This has important implications because changes in different genomic regions could have different indirect negative effects, and it will be difficult to predict these based on trait observations alone if such traits changes can have contrasting genomic underpinnings. On the positive side, if there are many genomic solutions to getting smaller, fish populations are less likely to lose all the genetic variation that would enable reversal towards getting large again if fishing stops. These multiple genomic solutions could therefore in some cases facilitate reversal of human-induced evolution, even if this reversal may be slow or incomplete. In addition to these insights, the project has provided important methodological development by illustrating an efficient and practical workflow that can to applied to obtain whole genome sequence data for hundreds of individuals of any species at a relatively low cost. It has also generated the first transcriptome assembly for the study species, the Atlantic silverside, and low-coverage whole genome sequencing for >850 individuals, all of which is publicly available and will form an important resource for future research. Last Modified: 04/01/2019 Submitted by: Stephen R Palumbi