Bees & Insects

Estimating variance components and breeding values in honeybee populations with pedigree or genomic information

---

Abstract

The aim of this work is to evaluate different phenotype, pedigree, and genomic data collection strategies and models for estimation of quantitative genetic parameters in honeybee populations. Breeding programmes in honeybees are gaining in number and importance. However, the biological peculiarities of honeybees are preventing the use of standard models to estimate quantitative genetic parameters. Due to polyandrious mating, the colony is a genetically diverse group of individuals contributing to a single colony phenotype value, which departs from the standard quantitative genetic models. Previous research developed methods to estimate breeding values using the colony-level phenotype values by decomposing the genetic variation between colonies into breeding values of queens, worker groups, and father groups. The developed method was tailored to a specific setting of controlled mating stations with a known pedigree for the queens. However, we might not have a complete pedigree for all the phenotyped colonies where some queens are open mated with drones. In such cases, we can use genomic information as a realised relatedness to improve estimation of quantitative genetic parameters using all the collected phenotypic values. To facilitate such scenarios, we have developed and tested models where the phenotype information is not on the same level as the genotype information, since the first one would be collected on a colony and the second one on (a group of) individuals. To obtain phenotypic data, we simulated a yearly cycle of a honeybee population with the SIMplyBee R package. We simulated the phenotype value as the sum of genetic effect, non-genetic effects, and a residual. We created a set of scenarios that first varied the level of the phenotype. In applications, phenotype is only measured on a colony level as the combined effort of multiple individuals. But the simulation enables us to also obtain queen and worker phenotypes. Second, we varied the genotyping strategy by collecting i) only queen genotypes; ii) pooled genotypes of n workers per colony with varying n; and iii) queen and pooled workers genotypes. In these scenarios, the phenotype level either matched the genotype level or not. And last, we implemented scenarios with and without considering the spatial variation. This is an important source of environmental variation in honeybee performance because they live and forage in local environments. We simulated continuous spatial variation with the R package INLA that was added to the phenotype simulation as well as to the estimation models as a random effect. We fitted all models in a full Bayesian framework using INLA method. Our results shed light into the sampling scheme required to reliably estimate quantitative genetic parameters using phenotype, pedigree, and genomic data in honeybee population and will therefore guide future research as well as breeding efforts.

Keywords: 2026

How to Cite:

Pustovrh, N., Gorjanc, G. & Obšteter, J., (2026) “Estimating variance components and breeding values in honeybee populations with pedigree or genomic information”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286302. doi: https://doi.org/10.31274/wcgalp.23913

Rights: 1