Aaron Shafer is an Assistant Professor at Trent University, Canada. Their research uses genomic and bioinformatics tools to characterize adaptive and demographic processes in natural populations. Research organisms in the lab currently include shrews, deer, caribou and mountain goats. Lead-author Aidan Jamieson was an honours biology student who is now doing a MSc at York University on bee genomics. As a lab we are committed to eradicating systemic racism in science and beyond and support BLM and #SHUTDOWNSTEM movements.
White-tailed deer are a billion dollar species. Ranging from real estate-prices, to outdoor gear, to hunting tags, the economic impact of this big-game species is wide-spread. Moreover, both hunters and farmers like big bodied and antlered deer. This preference, in quantitative genetics terms, has the potential to create a response to selection if the selected traits (like antlers) are heritable (h2). Anecdotally, many game managers and farmers will argue food and environment are the main drivers, but a handful of captive studies have shown many of these desired traits to be heritable (Williams et al. 1994; Lukefahr and Jacobson 1998; Michel et al. 2016).
We were interested in quantifying heritability of antler and body size traits in a free-ranging population of deer; but why do this if we already know the traits are heritable? There are two main reasons: i) the availability of genomic tools means we can estimate heritability without the need for pedigrees; and ii) heritability estimates in captive populations might not be reflective of wild populations for a variety of reasons. Candidly, lead author Aidan Jamieson also needed an undergraduate thesis project and we thought building genomic relatedness matrices and running mixed-effects models was appropriate for an eight-month undergraduate thesis…
What did Aidan find in his recently published undergraduate thesis? Body size metrics like muscle mass and hind foot length were highly heritable (h2> 0.50); antler metrics like number of points and spread showed lower values (h2 0.07-0.33). Our estimates were generally consistent with previous work on captive studies, but perhaps most relevant to the broader community was how filtering the genomic data influenced h2estimates [spoiler: filtering data has a large influence!]: we provided some suggestions on how to treat your data, but you will have to read the paper for this (see Jamieson et al. 2020). Another interesting aspect buried in supplemental was including year – a proxy for annual environmental variation – in the model, did not appreciably change the coefficients suggesting other more fine-scale environmental factors are at play. Ultimately, what this means is such traits might respond to selection, which we argue is most likely to be the case in farmed or small and isolated populations.
References