Introduction

The Western honey bee (Apis mellifera , hereafter honey bee), is a key pollinator of agricultural crops (Klein et al., 2007). To date, more than 27 subpopulations have been reported globally, which can be grouped into four distinct lineages, namely M (Western and Northern Europe), C (Eastern Europe), O (Near East and Central Asia) and A (Africa) (Cridland, Tsutsui, & Ramírez, 2017; Friedrich Ruttner, 1988). These lineages are characterized by differences in morphology, physiology and behaviour (Friedrich Ruttner, 1988). Within their native range, honey bees are commonly kept in hives for honey production and pollination purposes. In Europe, several selection programmes have been initiated to increase the productivity of honey bees (Adam, 1983; Büchler, Berg, & Le Conte, 2010; Chauzat et al., 2013; Guichard et al., 2020; H. Ruttner, 1972; Uzunov, Brascamp, & Büchler, 2017), while in Africa the majority of honey bees essentially evolved under natural selection (Dietemann, Pirk, & Crewe, 2009).
In some countries, honey bees are considered as a domesticated livestock species (Geldmann & González-Varo, 2018), due to the strong impact of human-mediated selection. In the beginning of the 19thcentury, importation of foreign honey bees among European regions began to increase, which profoundly reshaped the genetic structure of this species (Parejo, Wragg, Henriques, Charrière, & Estonba, 2020). Historically, native honey bees of Europe mainly belong to M and C evolutionary lineages. They are locally adapted to different climatic and geographical regions, resulting in several subspecies (Momeni et al., 2021; Friedrich Ruttner, 1988). Nevertheless, beekeepers in Northern Europe continue to replace native honey bees (A. m. mellifera ) with South-European honey bees (e.g. A. m. carnicaand A. m. ligustica ), as these subspecies are known to be more productive, gentle and calm (Bouga et al., 2011; Guichard et al., 2021). This practice has led to multiple admixture events between subspecies and the extinction of locally adapted honey bees (Bieńkowska, Splitt, Węgrzynowicz, & Maciorowski, 2021; F Ruttner, 1995). Furthermore, native honey bees are threatened by the widespread use of stabilised hybrid strains such as Buckfast (Adam, 1983; Bieńkowska et al., 2021).
The relocation of subspecies accompanied by admixture is a major risk factor of losing local adaptation and genetic diversity of honey bees (De la Rúa, Jaffé, Dall’Olio, Muñoz, & Serrano, 2009). Therefore, several conservation programmes have been initiated to maintain the genetic diversity of native honey bees. In Switzerland, the first conservation area for A. m. mellifera was established in 1977 in canton Glarus (Soland-Reckeweg, Heckel, Neumann, Fluri, & Excoffier, 2009 ). Nowadays, an additional conservation area exists in canton Obwalden. The two conservatories encompass a total area of 830 km2 and ~1050 colonies (Parejo et al., 2016). To limit admixture events with other foreign subspecies (e.g.A. m. carnica and Buckfast) these areas are typically located in remote alpine valleys. Besides the maintenance of the conservation areas, the breeding association of A. m. mellifera (mellifera.ch) established a selection programme including several mating stations. These stations are also located at geographically isolated areas and consist of 10 up to 20 selected drone-producing colonies. Currently, an ancestry-informative (microsatellites or single nucleotide polymorphisms; SNPs) marker panel is applied to determine the admixture level of conserved and selected A. m. mellifera colonies, and highly admixed colonies (>10%) are replaced with purebredA. m. mellifera (Parejo, Henriques, Pinto, Soland-Reckeweg, & Neuditschko, 2018). However, the replacement of admixed queens is expected to lead to an increase in inbreeding that could be detrimental to the small conserved A. m. mellifera population. Given that the survival of honey bees is strongly dependent on the genetic diversity (Jones, Myerscough, Graham, & Oldroyd, 2004; Kryger, 1990; Mattila, Rios, Walker-Sperling, Roeselers, & Newton, 2012; Mattila & Seeley, 2014; Oldroyd, Rinderer, Harbo, & Buco, 1992), monitoring of inbreeding in small conserved populations, such as A. m. mellifera in Switzerland, is crucial.
Inbreeding level indicates the probability that an animal receives the same allele from both parents. Genetic marker information allows to determine that alleles are identical-by-descent (IBD), while pedigree-based estimations require prior knowledge of individual ancestry (Kardos, Luikart, & Allendorf, 2015), which in case of the honey bee is often not available. Runs of homozygosity (ROH), IBD transmitted haplotypes, are one of the tools to estimate inbreeding levels without ancestry information (McQuillan et al., 2008). In fact, the length of ROH segments can be used to ascertain historical changes in population size and structure including admixture (few and short ROH segments), current inbreeding (multiple and long ROH segments) and a recent bottleneck (multiple and short ROH segments); see Ceballos (2018) for a complete review. Furthermore, it is possible to derive the genomic inbreeding coefficient (FROH) for an animal by dividing the sum of all homozygous segments (SROH) by the length of the analysed genome (McQuillan et al., 2008). Numerous studies have demonstrated that overlapping ROH segments, so-called homozygosity islands, can be successfully used to identify selection signatures in cattle (Purfield, Berry, McParland, & Bradley, 2012; Zhang, Guldbrandtsen, Bosse, Lund, & Sahana, 2015), sheep (Mastrangelo et al., 2017; Purfield, McParland, Wall, & Berry, 2017; Signer‐Hasler, Burren, Ammann, Drögemüller, & Flury, 2019), and horses (Druml et al., 2018; Grilz-Seger, Druml, et al., 2019; Grilz-Seger et al., 2018; Metzger et al., 2015), as well as in cultivated plants such as avocados (Rubinstein et al., 2019), almonds (Pavan et al., 2021) and pears (Kumar et al., 2020).
In this study we investigated the utility of colony genotypes derived from pooled workers to identify ROH segments in honey bees. Furthermore, we integrated individual admixture and FROH in a high-resolution population structure analysis to enhance the genetic monitoring of conserved A. m. mellifera. Finally, we screened the genome for homozygosity islands to detect selection signatures betweenA. m. mellifera and A. m. carnica honey bee colonies, related to geographic adaptations and human mediated selection.