PolyMaGNet : A Hybrid Method for Inferring Putative Causal Phenotypic Networks Accounting for Polygenic and Major Gene Effects

ABSTRACT Graphical models, such as Bayesian networks and structural equation models, have been widely used in quantitative genetics to investigate putative causal relationships among phenotypic traits. Constraint‐based algorithms have been proposed to search for recursive causal structures among phenotypes conditional on additive genetic effects; however, these approaches typically account only for the polygenic background and do not explicitly incorporate major gene effects that may also influence the traits under study. Here, we propose PolyMaGNet (Polygenic traits with Major Gene Network analysis), a hybrid method for inferring putative causal phenotypic networks while accounting for polygenic effects modelled through pedigree‐based additive relationships, together with major gene effects represented by markers associated with quantitative trait loci (QTL). The method combines a Bayesian multiple‐trait animal model, in which major loci are included as covariates, with the inductive causation algorithm applied to posterior samples of the residual covariance matrix. Candidate directed graphs are then compared using Akaike's information criterion to select a final fully oriented putative causal structure. In simulations based on a QTL mapping population, PolyMaGNet recovered the true skeleton more frequently than a method that conditions only on polygenic effects and selected the true fully oriented network in most simulation replicates. The method was further applied to an F2 Duroc × Pietrain pig resource population to reconstruct putative causal networks among longitudinal backfat, meat‐quality, fat‐deposition, and carcass‐composition traits. The results indicate that explicitly incorporating major gene effects can improve reconstruction of putative phenotypic networks among complex traits and provide a useful framework for investigating directional relationships in multivariate genetic architecture in animal breeding populations.

Authors

Institutions

Publication Details

Journal
Journal of Animal Breeding and Genetics
Published
2026-11-14
DOI
https://doi.org/10.1111/jbg.70087
Primary Topic
Genetic and phenotypic traits in livestock
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

PolyMaGNet : A Hybrid Method for Inferring Putative Causal Phenotypic Networks Accounting for Polygenic and Major Gene Effects

Renan Mercuri Pinto, Guilherme J. M. Rosa, Catherine W. Ernst, Fernando Brito Lopes et al.
Journal of Animal Breeding and Genetics
Genetic and phenotypic traits in livestock
article

PolyMaGNet : A Hybrid Method for Inferring Putative Causal Phenotypic Networks Accounting for Polygenic and Major Gene Effects

Renan Mercuri Pinto, Guilherme J. M. Rosa, Catherine W. Ernst, Fernando Brito Lopes, Bruno Dourado Valente, Mehdi Momen
article en

Abstract

ABSTRACT Graphical models, such as Bayesian networks and structural equation models, have been widely used in quantitative genetics to investigate putative causal relationships among phenotypic traits. Constraint‐based algorithms have been proposed to search for recursive causal structures among phenotypes conditional on additive genetic effects; however, these approaches typically account only for the polygenic background and do not explicitly incorporate major gene effects that may also influence the traits under study. Here, we propose PolyMaGNet (Polygenic traits with Major Gene Network analysis), a hybrid method for inferring putative causal phenotypic networks while accounting for polygenic effects modelled through pedigree‐based additive relationships, together with major gene effects represented by markers associated with quantitative trait loci (QTL). The method combines a Bayesian multiple‐trait animal model, in which major loci are included as covariates, with the inductive causation algorithm applied to posterior samples of the residual covariance matrix. Candidate directed graphs are then compared using Akaike's information criterion to select a final fully oriented putative causal structure. In simulations based on a QTL mapping population, PolyMaGNet recovered the true skeleton more frequently than a method that conditions only on polygenic effects and selected the true fully oriented network in most simulation replicates. The method was further applied to an F2 Duroc × Pietrain pig resource population to reconstruct putative causal networks among longitudinal backfat, meat‐quality, fat‐deposition, and carcass‐composition traits. The results indicate that explicitly incorporating major gene effects can improve reconstruction of putative phenotypic networks among complex traits and provide a useful framework for investigating directional relationships in multivariate genetic architecture in animal breeding populations.

Journal of Animal Breeding and Genetics
University of Wisconsin–Madison (US), Auburn University (US), Michigan State University (US)
Genetic and phenotypic traits in livestock
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon