AI-Driven Genomic Selection Models for Multi-Disease Resistance in Small Ruminants

AI-Driven Genomic and Immunogenetic Selection Models for Multi-Disease Resistance in Small Ruminants**

Background and Rationale
Small ruminant production systems face persistent and economically significant health challenges including paratuberculosis (Johne’s disease) caused by Mycobacterium avium subsp. paratuberculosis (MAP), caseous lymphadenitis (CLA) caused by Corynebacterium pseudotuberculosis, and neonatal diarrhea, a major cause of early lamb mortality often linked to inadequate passive immunity. These diseases reduce productivity, increase culling and mortality rates, and impose long-term management and diagnostic costs. Substantial variation in individual immune responsiveness across flocks suggests a strong genetic and immunogenetic basis for disease resilience. However, current breeding programs rarely incorporate immune biomarkers or genomic estimates of disease resistance into selection decisions.

Overall Aim
This project aims to develop machine learning–supported genomic and immunogenetic selection models that enable data-driven, multi-disease resistance breeding in small ruminants. By integrating high-density SNP genotyping, immunological biomarkers, serological assays, and environmental/farm-level risk factors, we aim to build predictive tools that enhance flock resilience and reduce disease burden in a sustainable, One Health-aligned manner.

Specific Objectives

1. Genomic Characterisation: Conduct GWAS and genomic profiling to identify SNPs, QTLs, and genomic regions associated with resistance to MAP, CLA, and variable maternal antitoxin responses.

2. Immunogenetic Profiling: Analyse innate and adaptive immune pathways (Th1/Th17, macrophage activation, antibody kinetics, cytokine signatures) influencing susceptibility and post-vaccination immune response.

3. Machine Learning Modelling: Develop and validate multi-disease predictive models (Random Forest, XGBoost, SVM, artificial neural networks) integrating genomic, immunological, and environmental data to classify animals as resistant, intermediate, or susceptible.

4. Genomic Selection Tools: Generate genomic EBVs (gEBVs) and construct a multi-disease resistance index incorporating production traits and resilience scores.

5. Decision-Support Platform: Develop breeder-friendly digital tools for real-time estimation of animals’ resistance potential, supporting practical selection and management strategies.

Innovation
The project is innovative in unifying three major disease challenges into a single AI-driven, multi-trait selection framework. It integrates field phenotyping (abscess scoring, MAP diagnostics, lamb diarrhea monitoring), immune responsiveness (ELISA titers, cytokine profiles), high-density genomic data, and ML algorithms to develop the next generation of precision breeding tools. This approach goes beyond traditional genetic evaluations by incorporating complex immunological and farm-level variables into predictive breeding decisions.

Expected Impact

• Enhanced flock-level resilience through genomic selection for immune competence.

• Reduced prevalence and economic losses resulting from MAP and CLA.

• Improved lamb survival through stronger maternal antibody responses and colostrum-mediated immunity.

• Contribution to antimicrobial reduction strategies in line with EU One Health objectives.

• Establishment of an international, multidisciplinary research network and preparation for Horizon Europe, BBSRC, TÜBİTAK, and other funding opportunities.

Collaboration Needs
The project seeks partners in quantitative genetics, ruminant immunology, veterinary microbiology, diagnostic development, epidemiology, bioinformatics, artificial intelligence, and sustainable livestock systems. Collaboration may include joint field studies, data integration, model development, cross-validation, and co-preparation of international proposals.