A genomic prediction framework for growth traits in black soldier fly (Hermetia illucens) breeding programmes
- 1James Cook University, Australia
Abstract
Background Genomic prediction has been applied in livestock, plant and aquaculture species to accelerate genetic gain. However, its application and potential still remain unexplored in farmed black soldier fly (BSF), an emerging globally important insect production species. Due to the diversity of BSF feed sources, selection strategies need to be evaluated across various feedstocks and production environments to aid in building genomic selection into BSF breeding programmes. In this study, 2,083 BSF larvae from the base population were reared on three diets (SYK: Soy-okara, BSG: brewers' spent grain and FVW: fruit and vegetable waste), and 536 from the F7 population were reared on a chicken starter feed diet. Larvae were phenotyped for four growth traits (LBW, larval body weight; LL, larval length; LW, larval width; LSA, larval surface area) and genotyped across 4,680 SNPs. Genomic prediction was evaluated using seven methods under additive (A) and additive-dominance (AD) models, considering across- and within-diet effects, genotype by diet (G x D) interactions, differences amongst reference and test generations, and SNP densities. Results Performance varied across all scenarios, with LBW showing higher predictive ability (PA) among the growth traits. Accounting for dominance generally improved PA. Multivariate models accounting for G x D performed similarly to, or marginally better than, within-diet predictions. Cross-diet prediction was more variable and generally lower than within-diet prediction, showing reduced transferability across untested nutritional environments. Reducing the SNP density from 4,680 to 3,000 SNPs maintained comparable PA; however, PA decreased below 3,000 SNPs. Across multiple generations, predictive performance declined, resulting from reduced genetic connectivity and possible genotype-by-environment (G x E) interactions between the reference and testing populations. Conclusions Overall, these results provide an application-based framework for genomic prediction in BSF in semi- or commercial production settings. Results also highlight the importance of the genetic model, representative reference populations, and appropriate marker densities for genomic selection.
Related
- Genetic parameters and pedigree–based selective breeding outcomes for productivity traits in industrial black soldier fly (hermetia illucens L.) production systems
- Environmental Frass Remodels the Larval Gut Microbiome to Enable Black Soldier Fly (Hermetia illucens) Bioconversion of Livestock Manure
- Long-term storage of Hermetia illucens (Diptera: Stratiomyidae) larvae
- First come, first served: sperm plug-mediated sperm precedence in the black soldier fly (Hermetia illucens)
- Inbreeding without consequences? Limited effects of inbreeding on life-history and performance traits in Hermetia illucens
BibTeX
@misc{gowda2026genomic,
title = {{A genomic prediction framework for growth traits in black soldier fly (Hermetia illucens) breeding programmes}},
author = {K. B. Gowda and C. Massault and D. R. Jerry and K. R. Zenger},
howpublished = {bioRxiv},
year = {2026},
month = nov,
doi = {10.64898/2026.11.05.764768},
url = {https://www.biorxiv.org/content/10.64898/2026.11.05.764768v1},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - A genomic prediction framework for growth traits in black soldier fly (Hermetia illucens) breeding programmes
AU - K. B. Gowda
AU - C. Massault
AU - D. R. Jerry
AU - K. R. Zenger
PY - 2026
DA - 2026/11/05
PB - bioRxiv
DO - 10.64898/2026.11.05.764768
UR - https://www.biorxiv.org/content/10.64898/2026.11.05.764768v1
AB - Background Genomic prediction has been applied in livestock, plant and aquaculture species to accelerate genetic gain. However, its application and potential still remain unexplored in farmed black soldier fly (BSF), an emerging globally important insect production species. Due to the diversity of BSF feed sources, selection strategies need to be evaluated across various feedstocks and production environments to aid in building genomic selection into BSF breeding programmes. In this study, 2,083 BSF larvae from the base population were reared on three diets (SYK: Soy-okara, BSG: brewers' spent grain and FVW: fruit and vegetable waste), and 536 from the F7 population were reared on a chicken starter feed diet. Larvae were phenotyped for four growth traits (LBW, larval body weight; LL, larval length; LW, larval width; LSA, larval surface area) and genotyped across 4,680 SNPs. Genomic prediction was evaluated using seven methods under additive (A) and additive-dominance (AD) models, considering across- and within-diet effects, genotype by diet (G x D) interactions, differences amongst reference and test generations, and SNP densities. Results Performance varied across all scenarios, with LBW showing higher predictive ability (PA) among the growth traits. Accounting for dominance generally improved PA. Multivariate models accounting for G x D performed similarly to, or marginally better than, within-diet predictions. Cross-diet prediction was more variable and generally lower than within-diet prediction, showing reduced transferability across untested nutritional environments. Reducing the SNP density from 4,680 to 3,000 SNPs maintained comparable PA; however, PA decreased below 3,000 SNPs. Across multiple generations, predictive performance declined, resulting from reduced genetic connectivity and possible genotype-by-environment (G x E) interactions between the reference and testing populations. Conclusions Overall, these results provide an application-based framework for genomic prediction in BSF in semi- or commercial production settings. Results also highlight the importance of the genetic model, representative reference populations, and appropriate marker densities for genomic selection.
ER -