Bees & Insects

Targeted within-family phenotyping increases selection response in communal black soldier fly (Hermetia illucens) breeding: a simulation study

---

Abstract

Breeding of communal organisms such as black soldier fly (BSF) is usually based on tray or family records, so programs rely on among-family selection to avoid bias from common-environment (c2) effects. This is robust, but it leaves most Mendelian sampling variance unused, and in small nuclei, makes inbreeding control critical. Recent vision/AI tools now make it realistic to phenotype only a few individuals per family, creating scope to add a light within-family step to an otherwise among-family scheme. The objective was to quantify (i) how much extra response is obtained when families are first ranked on family-level estimated breeding values (EBV) and only a small subset within retained families is phenotyped and selected, and (ii) how family-level optimal contribution selection (OCS) changes the gain-inbreeding-variance balance under fixed BSF mating logistics. We simulated a 50-generation BSF breeding program from three founder sources and three pleiotropic traits of moderate heritability and partly antagonistic genetic correlations, aggregated in one economic index. Each generation, EBV were estimated from tray/family means, the best families were retained, and 70 males and 140 females were selelcted and mated at a 1:2 ratio; progeny from each mating were split across three trays to impose the intended c2 structure. Within those families we compared: (1) among-family selection only (sires and dams selected at random); (2) among-family selection plus sampled males (a small male subset per family phenotyped and ranked on deviation from the family mean); and (3) among-family selection plus sampled males and females (same step in both sexes). All schemes were run with and without family-level OCS at a coancestry target of 0.005. Relative to among-family selection only, adding the male step increased the final-generation true breeding value (TBV) index mean by about 10%, and doing so for both sexes increased it by about 14%, showing that phenotyping only a few within-family deviates can recover Mendelian sampling even when selection is triggered at family level. This additional response was accompanied by lower average additive variances across the three traits in the non-OCS runs (roughly halved when only males were sampled and reduced by about two-thirds when both sexes were sampled) and pedigree inbreeding of ≈0.01-0.011 per generation. Introducing family-level OCS changed TBV means only slightly (about 0.2-1% above the corresponding non-OCS runs) but reduced pedigree inbreeding per generation by about 9-22% and increased last-generation additive variance by about 5%, 17%, and 37% in the three schemes, showing that a modest coancestry constraint can slow variance erosion in BSF nuclei. Overall, an EBV-based, three-step, AI-enabled scheme that ranks families, samples a few within-family deviates, and applies family-level OCS can raise gain and maintain diversity, offering a practical pathway to more sustainable BSF breeding programs.

Keywords: 2026

How to Cite:

Musa, A., Reyer, H., Wimmers, K., Daş, G., Mielenz, M., Metges, C., Reinsch, N. & Wittenburg, D., (2026) “Targeted within-family phenotyping increases selection response in communal black soldier fly (Hermetia illucens) breeding: a simulation study”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2285516. doi: https://doi.org/10.31274/wcgalp.23728

Rights: 1