Molecular Breeding Methods Using Approximate Bayesian Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional genomic prediction methods, such as GBLUP, fail to accurately predict complex traits influenced by non-linear genetic effects and genotype-by-environment interactions, limiting their effectiveness in breeding programs.
Innovation Solution
The method employs approximate Bayesian computation (ABC) with a biological model, such as a crop growth model, to estimate the effects of genotypic markers and link them with the model, allowing for simultaneous modeling of complex traits and their component traits, whether observed or unobserved, to predict trait performance and breeding values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional genomic prediction methods (GBLUP) are used, then the breeding program can be simplified and computation time reduced, but prediction accuracy for complex traits with non-linear genetic effects deteriorates
Solution Approach 1:
The patent segments the complex trait prediction problem into multiple components by introducing intermediate component traits that bridge the gap between genotypic markers and the complex target trait. This segmentation allows the system to model non-linear relationships step-by-step through a hierarchy of traits rather than attempting to model the complex trait directly, thereby improving prediction accuracy while managing computational complexity.
Solution Approach 2:
The patent introduces component traits as intermediary variables between the genotypic markers and the complex target trait. These intermediate traits serve as mediators that capture non-linear genetic effects and genotype-by-environment interactions, enabling more accurate predictions without requiring the full complexity of direct modeling of the target trait.
2Adaptability or versatility
If conventional genomic prediction methods are used, then the breeding program can be completed faster with simpler analysis, but the ability to handle non-linear genetic effects and genotype-by-environment interactions deteriorates
Solution Approach 1:
The patent performs preliminary actions by estimating the effects of genotypic markers on intermediate component traits before using these estimates to predict the complex target trait. This preliminary estimation of marker effects on component traits allows the system to pre-process and store the non-linear relationships, enabling faster subsequent predictions while maintaining the ability to handle complex genetic effects and environment interactions.
3Measurement precision
If more component traits are modeled simultaneously, then prediction accuracy improves, but computational time and model complexity increase
Solution Approach 1:
The patent segments the computational task by organizing component traits in a hierarchical structure where simpler component traits can be estimated independently before being used in more complex predictions. This segmentation allows the system to process multiple component traits efficiently by dividing the computational workload into manageable stages, improving prediction accuracy while controlling computational time requirements.
Data Source
AI summary
Improved molecular breeding methods include a method in which an association data set is developed by associating the phenotypes of a broad population of individuals with the individual genotypes. The association data set is used in conjunction with a growth model in order to select breeding pairs likely to generate offspring with one or more desirable traits.