Trans-species polymorphisms are variants shared by descent among species (Bryja, Galan, Charbonnel, & Cosson, 2006; Charlesworth, 2006). A few well-known systems show trans-species polymorphism. Examples include the ABO blood grouping (Ségurel et al., 2012) and the HLA/MHC histocompatibility in primates (van Oosterhout, 2009). Trans-species variation predates speciation, and a common inference is that some form of balancing selection maintains the polymorphism. Thus, balancing selection is implicated in immune genes among cod species that show signs of phylogenetic inconsistencies of trans-species polymorphism (Halldorsdottir & Arnason, 2015). Polymorphic inversions that contribute to adaptive color pattern variation in Heliconius butterfly species are maintained through frequency dependent selection (Jay et al., 2018). Similarly, in deer mice a profusion of inversion polymorphism is likely maintained by divergent selection involved in local adaptation (Harringmeyer & Hoekstra, 2022). Inversions are important in genome evolution since they suppress recombination in heterozygotes (Sturtevant, 1917), preventing recombinational breakup of adaptive gene complexes. They can be difficult to detect using both long- and short-read sequencing technologies (M. J. Chaisson, Wilson, & Eichler, 2015; M. J. P. Chaisson et al., 2019) because inversion breakpoints are typically embedded within highly identical duplications (Kidd et al., 2008; Kidd et al., 2010; Sanders et al., 2016). Principal component analysis (Berg et al., 2016) and FST scans (Hemmer-Hansen et al., 2013), between populations from different locations, have previously been used to detect inversions in the Atlantic cod (Gadus morhua). Both methods are promising although long-read sequencing would be ideal for detecting structural genetic variation due to the cryptic breakpoints.
Several loci associated with haemoglobin and temperature have shown divergence in Atlantic cod populations (Borza, Stone, Gamperl, & Bowman, 2009; Bradbury et al., 2010; Case, Hutchinson, Hauser, Van Oosterhout, & Carvalho, 2005; Nielsen, Hansen, & Meldrup, 2006; Pogson, 2001; Sick, 1965). These loci showed characteristics of large chromosomal inversions such as linkage disequilibrium (Berg et al., 2016; Bradbury et al., 2010; Hemmer-Hansen et al., 2013; Kirubakaran et al., 2016; Sodeland et al., 2016) and later confirmed with long-read sequencing techniques on chromosomes 1, 2, 7, and 12 (Kirubakaran et al., 2020). Selection affects these inversions as the frequency varies between ecotypes and they are maintained by stabilizing selection on both sides of the Atlantic (Matschiner et al., 2022; Sodeland et al., 2022). The most striking result is that we find evidence for inversions in the polar cod (Boreogadus saida) in the same or similar chromosomal locations that are known in the Atlantic cod (Gadus morhua) on chromosome 1 and 7 (Berg et al., 2016; Kirubakaran et al., 2016). These inversions are to some extent enigmatic in the Atlantic cod by showing cryptic variation with different spatiotemporal origins (Bradbury et al., 2014; Kess et al., 2020; Pogson & Mesa, 2004; Refoyo-Martínez et al., 2019).
The polar cod, a relative of Atlantic cod, is an important species in the Arctic fauna. Despite its importance science has neglected this species. It shares two common names with the species Arctogadus glacialis, Arctic cod and polar cod. Conventionally it is called Arctic cod in America and polar cod east of the Atlantic. Here, we follow the eastern tradition. Welch et al. (1992) state that the polar cod transfers 75% of energy from the plankton to upper trophic level species. It comprises more than half of the narwhal (Monodon monoceros) diet and is the predominant prey of beluga whales (Delphinapterus leucas) (Loseto et al., 2009; Welch et al., 1992). It is adapted to the extreme Arctic environment. This is shown, for example, by the fact that antifreeze glycoproteins constitute the major fraction of protein in the blood serum of the polar cod (Harding, Anderberg, & Haymet, 2003). Antifreeze glycoproteins suppress ice crystal growth and protect tissues from cold-induced damage (Chen, DeVries, & Cheng, 1997; Geoghegan, Osuga, Ahmed, Yeh, & Feeney, 1980; Harding et al., 2003; Scholander, Van Dam, Kanwisher, Hammel, & Gordon, 1957). Being stenothermal and showing little acclimation potential (Leo, Graeve, Storch, Portner, & Mark, 2020), recent warming has resulted in distributional shifts and thus reduced the availability of polar cod to marine birds. This is alarming as the polar cod is the primary forage fish of marine birds in the Arctic and for the Arctic seal species (Divoky, Lukacs, & Druckenmiller, 2015; Welch et al., 1992). Because of its sensitivity to temperature, polar cod could be used as a monitoring instrument of global climate changes in the Arctic ecosystem. Huserbraten, Eriksen, Gjosaeter, and Vikebo (2019) and Gjosaeter, Huserbraten, Vikebo, and Eriksen (2020) have made attempts in doing so by analyzing past observations of the environment and the polar cod for future prediction models. Results suggest northward retreat and possible recruitment collapse if warming and ice-reduction continues. Observations show differences of prey composition, phenotypic traits, predation risk, and reproduction strategies between climatic domains around Svalbard Archipelago which is surrounded by Arctic, Barents, and Atlantic seas and is thus ideal for observing different populations of polar cod (Nahrgang et al., 2014). This gives hope for the survival of the polar cod and its response to environmental change. The habitat being circumpolar with few barriers, the polar cod has been considered panmictic with little population differentiation (Pálsson, Källman, Paulsen, & Árnason, 2008; Robert E. Wilson et al., 2017; R. E. Wilson et al., 2019). Recent analysis of microsatellite data has shown significant, albeit weak population differentiation (Gordeeva & Mishin, 2019; Madsen, Nelson, Fevolden, Christiansen, & Præbel, 2015; Maes et al., 2021; Nelson et al., 2020) and RADseq data displayed distinct genetic clustering unrelated to geographic locations (Quintela et al., 2021).
Here we present novel whole-genome sequence (WGS) results showing population differentiation in polar cod. Two spatially separated sampling localities are compared, Besselfjord and Tyrolerfjord from north-eastern Greenland (see Fig. 1). To detect signatures of inversion polymorphism, we used principal component analysis (PCA), differentiation measured by FST between populations, which are defined as geographically separate sampling localities, scans of the ABBA-BABA D-statistic (Durand, Patterson, Reich, & Slatkin, 2011; Green et al., 2010; Patterson et al., 2012), scans of neutrality statistics (Korneliussen, Moltke, Albrechtsen, & Nielsen, 2013), by using a new approach of scanning the percentage of variation explained by the first principal component of PCA in windows across a chromosome, and by local PCA analysis (Li & Ralph, 2019) scanning all chromosomes of the genome.