Identity-by-Function BLUP for Allelic Heterogeneity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting unobserved phenotypes and selecting genetic variants for genetic improvement in agricultural species and human genetics are limited in efficiency and accuracy, particularly due to inadequately capturing allelic heterogeneity and functional equivalence, leading to incomplete and inaccurate predictions.
Innovation Solution
A method involving the computation of a functional unit dosage matrix and an identity by function relationship matrix to predict phenotypes and select organisms with improved performance, using statistical models like linear regression and neural networks to aggregate functionally equivalent alleles, thereby improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional phenotype-based breeding or standard genomic selection is used, then the breeding process is simple and easy to implement, but the prediction accuracy of unobserved phenotypes and selection of genetic variants is limited
Solution Approach 1:
The patent segments the genome into functional units (genes, pathways, gene families) and groups alleles into functional equivalence classes based on their biological functions rather than treating all genetic variants equally. This segmentation allows the method to capture allelic heterogeneity by analyzing functionally equivalent alleles separately, thereby improving prediction accuracy for complex traits while maintaining a structured approach to genome-wide data analysis
Solution Approach 2:
The patent introduces functional units and identity-by-function relationship matrices as intermediary structures between raw genotype data and phenotype prediction. These intermediaries aggregate alleles based on functional equivalence and establish relationships among individuals through shared functional alleles, serving as a bridge that translates complex genomic data into improved predictive models without requiring direct analysis of every individual variant
2Reliability
If standard genomic relationship matrices based on identity by descent are used, then the method is computationally efficient, but it fails to capture allelic heterogeneity and functional equivalence
Solution Approach 1:
The patent changes the fundamental parameter for measuring genetic relationships from identity-by-descent (IBD) to identity-by-function (IBF). Instead of relying on shared ancestry and allele inheritance patterns, the method defines relationships based on sharing functionally equivalent alleles, transforming the relationship matrix construction approach to better reflect functional genomic equivalence and improve prediction reliability for traits influenced by allelic heterogeneity
3Measurement precision
If computational techniques and machine learning methods are applied to predict phenotypic consequences, then prediction capability is enhanced, but efficiency and accuracy of selecting genetic variants remain limited
Solution Approach 1:
The patent performs preliminary grouping of alleles into functional equivalence classes before conducting phenotype prediction and variant selection. By pre-organizing genetic variants based on functional equivalence and computing identity-by-function matrices in advance, the method reduces the complexity of subsequent analysis steps and improves selection accuracy by ensuring that functionally equivalent alleles are consistently treated together throughout the prediction process
Data Source
AI summary
Provided herein are methods for predicting unobserved phenotypes and selecting genetic variant organisms for effective use in genetically improving agricultural species, as well as in human genetics and medicine. Also provided herein are systems for implementing such methods, as well as computer-readable storage media storing instructions for performing such methods.