Genetic breeding and multi-omics integration for alfalfa improvement from trait discovery to cultivar development

Abstract

Alfalfa (Medicago sativa L.) is a globally cultivated perennial forage legume whose productivity is increasingly constrained by climate-induced stresses, yet its autotetraploid genome, self-incompatibility, and 8–12-year breeding cycles impede rapid cultivar development. This review critically evaluates the transition from conventional phenotypic selection to data-driven breeding strategies, examining genomic selection (GS), genome-wide association studies (GWAS), multi-omics integration, CRISPR/Cas9 genome editing, and high-throughput phenotyping (HTP) within the context of polyploid crop improvement. We demonstrate that GS offers the most practical near-term path for polygenic trait improvement, with prediction accuracies enhanced through marker importance weighting and genotype-by-environment covariance modeling, while GWAS and pan-genome analyses, including structural variant incorporation improving GS accuracy, enable high-resolution trait dissection. Multi-omics integration shows greatest utility for oligogenic traits with moderate-to-high heritability, whereas CRISPR/Cas9 has enabled functional validation of key loci but faces transformation bottlenecks and regulatory barriers precluding commercial deployment. HTP platforms coupled with machine learning provide scalable phenotyping, though standardization gaps and infrastructure costs persist. We propose an integrated digital breeding pipeline connecting trait discovery through cultivar deployment, supported by community reference resources and actionable breeding recommendations including rapid-cycle GS and sparse testing designs. This framework positions alfalfa breeding for systematic translation of molecular discovery into climate-resilient, high-yielding cultivars.