Haplotype bias detection using pedigree-based transmission simulation: traces of selection that occurred in apple breeding

TL;DR

This study uses pedigree-based simulations to detect selection in apple breeding, revealing genomic regions shaped by historical breeding practices.

Contribution

A novel method for detecting selection in apple breeding without requiring phenotypic data, using haplotype bias from pedigree simulations.

Findings

Biased haplotype regions overlapped with loci linked to fruit traits like fructose content.

Gene Ontology analysis showed enrichment for stress-related and developmental genes in selected regions.

The method can uncover cryptic selection signals and identify underutilized founder alleles for breeding improvement.

Abstract

Breeding perennial fruit trees like apple is constrained by long generation times and limited population sizes, which often lead to repeated use of a few elite cultivars and consequently narrow genetic diversity. To better understand how such selection processes have shaped the current genetic structure, we applied gene-drop simulations—a pedigree-based method using known parentage and genetic maps—to a curated set of 185 apple cultivars used in Japanese breeding programs, genotyped with 11 786 genome-wide single nucleotide polymorphism markers. This approach enabled us to quantify the expected distribution of founder haplotypes and identify genomic regions where observed founder haplotype frequencies significantly deviated from expectation, suggesting potential selection. Notably, biased regions overlapped with loci associated with key fruit traits, such as fructose content, exemplified by an increase in haplotypes from “Golden Delicious.” Furthermore, Gene Ontology analysis revealed enrichment for regions containing genes involved in stress-related and developmental functions, pointing to broader physiological traits under selection. Unlike traditional methods requiring phenotype data, our approach does not depend on trait measurements and can thus uncover cryptic selection signals, including traits that were not explicitly targeted during breeding. This method offers a framework for identifying overlooked genetic regions and underutilized founder alleles, which can be reintroduced to broaden the genetic base and improve breeding outcomes. Furthermore, the approach is adaptable to other perennial crops with available pedigree and genomic data. Our findings demonstrate the power of integrating pedigree structure with genomic information to reveal both historical and ongoing selection in structured breeding populations.

Figures