Introduction
Understanding how organisms adapt to their local environment is one of the central questions of evolutionary biology, which is becoming increasingly important in a world of human-induced rapid climate and environmental change (HIREC). It is generally accepted that genetic variation within and among populations is influenced by the local environment in which organisms reside. For example, populations along an environmental gradient may be adapted to their local conditions if selection is strong enough relative to drift and gene flow between populations (Kawecki & Ebert, 2004)⁠. Since local adaptation arises from natural selection on adaptive phenotypic traits, it can be demonstrated by genetic differentiation at the genetic loci underlying those traits (Phifer-Rixey et al., 2018; Stillwell, 2010; Stinchcombe et al., 2004). The genetic basis for environmental adaptation has been uncovered for a few obvious traits with distinct phenotypic characteristics, such as variation in coat colour in mice in relation to environmental background colour (Linnen et al., 2009; Nachman et al., 2003) or reduction in armour plating in sticklebacks in response to freshwater colonization (Colosimo et al., 2005; Cresko et al., 2004)⁠, or body size and blood chemistry (Phifer-Rixey et al., 2018).
Theoretical and empirical studies suggest that many adaptive processes have a polygenic basis and are controlled by many genes of small effect (Barghi et al., 2020; Yeaman, 2015). Conventional genome scans are good at detecting signals from adaptive loci with large effects, but their ability to detect weak signals of polygenic selection acting across many loci is rather limited, because they focus on only one locus at a time (Rellstab et al., 2015; Wellenreuther & Hansson, 2016). In contrast, multivariate genotype-environment association (GEA) methods offer the possibility of analysing many loci simultaneously together with several environmental predictors. These are therefore suitable for detecting signals of polygenic selection (Capblancq et al., 2018; Forester et al., 2018),⁠ but have not yet been widely used in evolutionary ecology. In addition, these multivariate approaches also allow quantification of spatial patterns of adaptive genetic variation associated with environmental variables (Lasky et al., 2012; Micheletti et al., 2018; Nadeau et al., 2016)⁠.
Small forest mammals provide an ideal biological model to investigate the relative roles of selective and neutral factors in response to clinal environmental gradients, because they have large geographic ranges and individuals are not highly mobile within the range (Haasl & Payseur, 2016). Such gradients, and their respective genetic responses, include climate, and divergence in gene regulatory regions and genes related to metabolism and immunity (Phifer-Rixey et al., 2018) or body size and extremities ratio (Ballinger & Nachman, 2022); rural-urban gradients, and signals of selection in genes involved in lipids and carbohydrates metabolism (S. E. Harris & Munshi-South, 2017)⁠; as well as altitudinal gradients, and genes related to metabolic function and oxygen transport (Beckman et al., 2022; Waterhouse et al., 2018).
The bank vole Myodes glareolus (also known asClethrionomys glareolus ; Kryštufek et al., 2020) is a small Eurasian forest-dwelling rodent with a broad geographic distribution in Europe, ranging from the Mediterranean peninsulas and the southern coast of the Black Sea in the south almost to the northern edge of Scandinavia (Figure 1). This distribution covers a wide temperature gradient (Figure 1). Bank voles survived in several refugia during the Last Glacial Maximum, including the well known refugia on the Mediterranean peninsulas and cryptic refugia in the Carpathians (Kotlik et al., 2006; Wójcik et al., 2010) and the Ural Mountains (Abramson et al., 2009; Deffontaine et al., 2005)⁠. Their subsequent recolonization of Europe, when the climate became more favourable at the beginning of the Holocene, resulted in a complex genetic structure, with several distinct phylogeographic lineages, first described based on mitochondrial DNA sequences (Filipi et al., 2015) and later confirmed by genome-wide SNP analyses (Horniková et al., 2021; Marková et al., 2020). Bank voles have limited dispersal capabilities (Deter et al., 2008; Viitala et al., 1994) and short generation times, resulting in large local effective population sizes. Together with their limited capacity to disperse, these factors result in a large evolutionary potential for genetic responses to local conditions. Therefore, bank voles are a suitable system to study signatures of local adaptation in response to spatially varying climate-induced selective pressures along an environmental gradient. They have been the target of GEA-studies in relation to geographic expansion (White et al., 2013) and tolerance toPuumala orthohantavirus infection (Rohfritsch et al., 2018)⁠. However, the specific selection forces driving adaptations, as well as the genetic loci involved, are not well understood for wide latitudinal gradients.