Discussion
The maintenance and evolution of colour polymorphism remains one of the most appealing yet elusive evolutionary questions. Since diversity in life history strategies and complex traits have shown to segregate with colour-defined phenotypes, it is plausible to expect evidence of those associations also at a molecular level. Here we report the first assemblies of the tawny owl genome and a comprehensive genome-wide-association study leading to the discovery of structural variation associated with colour polymorphism – though not on the traditional M1CR gene.
Suggestion of a molecular basis of food intake, starvation response and energy homeostasis
The most obvious target for the genetic regulation of coloration in melanin-based pigmentation is the coding sequence of the geneMC1R . Contrary to general expectations, comparisons between coding sequences of grey and brown tawny owl revealed MC1R to be 100% conserved between morphs. These observations are however not exceptional, as other avian species exhibiting equally remarkable colour polymorphisms also show conservation of MC1R coding region (Avilés et al., 2019; Hoffman et al., 2014; MacDougall-Shackleton et al., 2003). Among the investigated genes that are to some extent involved in the melanin-production pathway, only the melanin concentration hormone receptor (MCHR ) exhibits non-synonymous substitutions between colour morphs. A direct and functional relationship between MCHR polymorphism and melanin pigmentation has been found exclusively in teleost fishes, specifically trout and salmonids (Diniz & Bittencourt, 2019). Among mammal and avian taxa alike, the functionality of this neuromodulator is rather associated with fasting control and energy balance, physiological traits plausible to be under selection in food limitation conditions (Cui et al., 2017).
In the genome-wide association study (GWAS) we expanded the molecular screen to an original dataset of 370 individuals and respective pedigrees. Correcting for both genetic relationship and population substructure the GWAS suggested additional evidence for a scenario where resource-related physiological responses are colour-morph specific, as predicted by Karell et al (2011). Associations with colour phenotypes revealed two candidate loci, FAM135A and FTCD, whose biological functions were shown to be linked to the lipid metabolism pathway, fat storage mechanisms, and maintenance of homeostasis under starvation in both pigs and chickens in captivity (Poleti et al., 2018; Zhang et al., 2021). We propose that in our tawny owl population the putative biological functionality of these candidate loci are better contextualized in the harsh winter conditions in Southern Finland. At 60° N latitude, Finnish tawny owls experience snowfall periods that usually start in December and generally peaking in January and February with snow-coverage usually lasting until the months prior to the onset of the tawny owl’s breeding season (Karell et al., 2011). Previous work on the study population used here have shown that these extreme conditions do impose a strong selective pressure against brown tawny owls while resulting in a relatively higher survival rate of grey tawny owls following snow-heavy winters (Karell et al., 2011). Since we find a big effect size of these candidate loci in predicting grey coloration (up to 100% in some genotype combinations) our findings suggest a genetic regulatory mechanism for grey coloration in this population as well as some of the loci underlying the expected adaptation to their local environment.
The relevance of an adapted lipid metabolism (whether it would be degradation, accumulation, or deposition) to survival in cold environments has been reported multiple times across taxa (Blem, 1976; Lucassen, Koschnick, Eckerle, & Portner, 2006). To the best of our knowledge, this study is unique in demonstrating a genetic link between lipid metabolism and melanin-pigmentation in a natural population, effectively rendering the tawny owl system as the first where the molecular basis of intraspecific melanin-associated phenotypes have been proposed.
A brief exploration of signatures of selection – via HWE tests and temporal shifts of candidate loci’s genotype frequencies – showed that, when they did occur, deviations from HWE were associated with an excess of homozygotes for the most common allele in either locus implying that selection appears to be against grey-coated individuals. Because ecological field studies rather reported a strong selection against brown individuals in harsh winter conditions, putative selection against grey morphs could be interpreted at the light of milder northern winters associated with ongoing climate warming(Räisänen, 2021).
Lighter plumage colour and putative trade-offs in heat absorption mechanisms
While the advantage of lighter plumage coloration in snow-covered environments might be expected under the background matching hypothesis, it is undeniable that less melanin pigmentation reduces the integument capacity to transform UV radiation into heat (Margalida, Negro, & Galván, 2008). Tawny owl genome analyses show that the architecture surrounding the candidate loci FTCD contains both a collagen (Col6a2 ) and a β-keratin gene (BISK1 ), both playing a key role in integument micro-structure (Kowata et al., 2014). Of particular relevance is the identification of BISK1 in the tawny owl genome. This novel copy of a β-keratin gene has only been recently identified in the chicken, and because its expression is confined to the barbule cells of that species’ contour feathers, it is suggested to be relevant for the formation of micro-barbules in pennaceous feathers – important insulation structures (Kowata et al., 2014). We hypothesize that the conservation of the Col6a2 -FTCD-B1SK1 region in tawny owls and the identification of a candidate loci associated with plumage colour also indicates the molecular basis of a trade-off, hypothetically linked to the grey phenotype. While there is empirical evidence for morphological differences between morphs where grey tawny owls have a higher number and density of barbules than brown morphs (Koskenpato et al., 2016), it is unknown whether this morphological difference in feather structure translates into morph-specific net insulative properties..