On Farm GS-Beans

On Farm GS-Beans - Alliance Bioversity International - CIAT

This project aims to determine if on-farm genomic selection can deliver a higher rate of genetic gain in farmers' fields, particularly in female-managed plots, than conventional breeding in which most selection is applied on research stations. 

Project Name: On Farm Genomic Selection-Beans 

Start and end date: 2022 to 2027 

Geographic area, countries: Tanzania 


FarmSelect-Beans will pilot a new, genomics-driven selection method based on on-farm performance of Stage 1 breeding material, enabling increased rates of genetic gain delivered to the farmers. We aim to generate GEBVs based on phenotyping on-farm, which will increase selection accuracy for performance under farmer management (particularly for female farmers) and enable rapid cycling of parents by reducing the number of years of testing before new crosses are made. The breeding lines will be evaluated for yield (using quantitative measurement) and farmer preference, with gender-intentional participatory methods to ensure that women’s preferences and experience as plot managers are adequately represented. We will use the Gender + product profile and the Gender-responsive participatory varietal selection tool to collect data on trait preferences along with the gendered implications of these traits to labour and inputs invested on farms.  The integration of plots managed by women, as well as the consideration of female household members’ preferences in GEBVs to select parents and advance candidate varieties, is a key element of this project and will significantly contribute to the design of target product profiles (TPPs). Their inclusion will increase the probability that new cultivars will be accepted by farmers. 

Key activities

  • Modelling and simulating on-farm GS programs to optimize pipeline design and model training. 
  • Estimating predicted gains on-farm for direct selection on-farm vs indirect selection on the research station using unselected Stage 1 lines. 
  • Conducting two cycles of on-farm and on-station recurrent GS 
  • Comparing genetic gains realized on farm from parallel recurrent selection program conducted with and without use of data from on-farm selection at Stage 1.