From Neglected to Next-Generation Crops: Breeding and Omics-Based Improvement of Underutilised Legumes
TL;DR
Evidence on the breeding and omics-based improvement of underutilised legumes is synthesised, identifying a persistent disconnection between genomic resource generation, downstream trait validation, breeding pipeline integration and farmer-level variety release.
Abstract
Underutilised legumes, a heterogeneous group encompassing Bambara groundnut, winged bean, moth bean, horse gram, grass pea, lablab, tepary bean, adzuki bean and rice bean, combine considerable environmental resilience and nutritional density with a long-standing shortage of genetic and genomic infrastructure. This review synthesises evidence on the breeding and omics-based improvement of these species, moving beyond a species-by-species inventory towards an integrated appraisal of methodological progress and remaining constraints. The literature indicates that reference genome assembly has advanced markedly for several species over the past decade, yet assembly contiguity, annotation completeness and population-scale resequencing coverage remain uneven across the group, with African orphan legumes generally trailing their Asian and American counterparts in genomic infrastructure. Genome-wide association and linkage mapping studies have identified marker-trait associations for drought tolerance, flowering time, seed composition and yield components, although modest population sizes, restricted environmental replication and an absence of independent validation frequently limit confidence in candidate gene assignments. Genomic selection, speed breeding and genome editing offer plausible routes to acceleration, but their application to underutilised legumes remains largely exploratory, constrained by the scarcity of well-phenotyped training populations and, for several species, by inefficient transformation and regeneration protocols. Evidence on nutritional composition and antinutritional factors is comparatively abundant but methodologically heterogeneous, which complicates cross-species comparison. The review identifies a persistent disconnection between genomic resource generation, downstream trait validation, breeding pipeline integration and farmer-level variety release. Future priorities include coordinated multi-environment phenotyping networks, harmonised trait ontologies, pangenome development and de novo domestication strategies capable of converting existing genomic knowledge into deployable, climate-resilient cultivars for smallholder production systems.