As a new beekeeper, I never really considered much about the origins of honey bees. I knew they were not native to the United States, most are from Europe and the ones that sting in a swarm are called Africanized. Local beekeepers talk about ordering Italians or Carniolans to populate hives, and during a recent local beekeeper’s association presentation, Asian honey bees were mentioned. From where Apis mellifera, the Western honey bee arose, I did not know.
As it turns out, the origin of honey bees is a highly debated topic. Some say they arose from Asia; others say Africa. Recently, researchers from the University of California—Davis used short nucleotide polymorphisms (SNPs) and two sets of previously published whole genome data, included additional sequenced genomes and applied multiple computational methods to analyze honey bee population genetics. They published their conclusions in Genome Biology and Evolution.
Cridland, Tsutsui and Ramírez began by laying out what was known from earlier research: A. mellifera is one of ten species in the genus Apis and the majority of these are distributed across Asia; historically, A. mellifera was distributed across sub-Saharan Africa, Europe, parts of western Asia, and the Middle East before domestication and transportation by Europeans to wherever they settled around the world. In A. mellifera’s native range, there have been at least 26 subspecies identified that are distinct in geographical and morphological ways. These different subspecies have been organized into five different lineages: M (Western Europe from Spain to Norway), C (Central Europe, including Italy and Austria), O (Turkey and Jordan), A (Africa) and Y (northeastern Africa and the Middle East, including Saudia Arabia and Yemen). Despite having defined, distinctive behaviorial and morphological traits, many of these honey bee lineages live in close proximity, suggesting a potential for relatedness.
Researchers defined over 832,000 SNP sites in the A. mellifera genome in which most individuals from the two data sets (total of 149 honey bees) were able to be genotyped. The African lineage (A) had the highest mean nucleotide diversity and the most SNPs restricted to only that lineage. O and Y lineages had mean nucleotide diversity between A lineage and C and M lineages.
Examining how much genetic variance existing among each lineage compared to the total western honey bee population found the Y lineage had the greatest differentiation compared to the rest. In fact, the differentiation between the A lineage and the C, M and O lineages were lower than differentiation among C, M and O populations. This lower differentiation suggests that C, M and O lineages are more closely related to the A lineage, adding weight to the hypothesis that the African lineage is the source of the population for C, M and O populations (Europe and the Middle East). Interestingly, the O lineage has greater genetic variance compared to the Y lineage than the A lineage. Based on the distribution of allele frequencies and the gaps in intermediate- and high-frequency SNPs, the C and M lineages likely were the result of two different waves of honey bees that came to Europe from Africa.
Using a software analysis program to test admixture (interbreeding of previously isolated populations), the C and O lineages were grouped together because they were likely less differentiated. This clustering suggests these two populations share a common ancestor. The admixture testing also revealed Y lineage bees grouped with African bees, but not with the O lineage. This is an interesting result as Y and O are both Middle Eastern populations and are close geographically.
A pairwise analysis of dissimilarity at each SNP site revealed that Y group individuals were placed between A lineage and Apis cerana, the Asian honey bee from which A. mellifera diverged 6–25 million years ago. This result agrees with other published data that support the hypothesis that northwestern Africa or the Middle East is the origin of A. mellifera. This pairwise analysis was also consistent with M, C and O lineages having descended from African ancestors, a hypothesis proposed by another researcher in 2006.
Taking calculations of group effect on subset genetic variation along with the groupings that resulted from the admixture analysis, Cridland, Tsutsui and Ramírez stated their results indicate that C and O lineages are the result of a single migration event from Africa and once in place, the population subsequently diversified. Because the O lineage was placed between the C and A lineages, it was likely that A and O lineages share a more distant common ancestor than C and O lineages. In addition, the relationship between the Y and A lineages suggested these bees descended from a population of A. mellifera ancestral to all western honey bees.
Analysis of differences among the African A. mellifera samples (from Kenya, new genomic data used for this analysis, South Africa and Nigeria) did not produce any influence related to geography. Performing a similar analysis of the M lineage samples as a control (similar number samples and subspecies from a similar number of different countries) showed a small effect of partitioning bees into two subsets: Spain versus Northern Europe (Sweden, Norway, Poland). However, because there was little to no geographic effect correlated to differentiation, these analyses suggest there is a large amount of gene flow among the A and M populations.
Additional analysis of recent admixture among populations suggest interbreeding in central European populations (Austria, Germany and Italy), two M group populations (Sweden and Norway) and the Jordanian population (with contributions from Turkey, an O lineage population, and Y lineage). However, there was no interbreeding among the African populations.
The analysis done by Cridland, Tsutsui and Ramírez tried to sort out some of the conflicting conclusions from the two genome data sets previously analyzed separately by other researchers. The current work showed that despite living next to each other, some populations have not mixed recently, and the residue from initial divergence had confused earlier analyses about which lineages were the ancestors and which were more recent population mixtures. The array of computer analysis tools used in the Genome Biology and Evolution article with a larger combined set of data suggests an origin for A. mellifera as an ancestor to A and Y lineages. This narrows the geographic location to northeastern Africa and the Middle East. The European and Middle Eastern A. mellifera populations are derived from African lineages, and subsequent diversification gave rise to the C, O and M lineages. However, the total set of genomic data has less than 160 individuals, and the authors acknowledge that data quality affects placement of lineages, making discrimination among A. mellifera lineages subject to reinterpretation when more high-quality data can be added. Regardless, I look forward to a more detailed family tree for the Western honey bee.
Cridland, J.M., Tsutsui, N.D. and Ramírez, S.R. (2017)The complex demographic history and evolutionary origin of the western honey bee, Apis mellifera. Genome Biol. Evol. [Epub ahead of print]. doi: 10.1093/gbe/evx009.