Case Study: Construction of the Largest and Most Genetically Diverse Upland Cotton Graph Pan-Genome

2026 Early Breakthrough: The world’s largest and most genetically diverse upland cotton (Gossypium hirsutum) graph pan-genome has been constructed. BMKGENE collaborated in genome sequencing (PacBio long reads, Hi-C), de novo genome assembly and pan-genome analysis.

Published: 02 January 2026
Journal: Nature Genetics
Author: Yang, Z., Yang, Z., Gao, C. et al.
Title: Graph pan-genome illuminates evolutionary trajectories and agronomic trait architecture in allotetraploid cotton
DOI: https://doi.org/10.1038/s41588-025-02462-1
Institution: State Key Laboratory of Cotton Bio-breeding and Integrated Utilization, Institute of Cotton Research, Chinese Academy of Agricultural Sciences, Anyang, China

1. Background

As a critical global economic crop for textiles, upland cotton contributes over 95% of global fiber output. However, long-term intensive cultivation and artificial selection have caused severe genetic diversity loss and cultivar homogenization. Coupled with biotic stresses, there is an urgent need to mine elite genes from germplasm, while existing genomic resources are insufficient for in-depth cotton research and improvement.

2. Strategy

To address these issues, 107 representative materials (covering 7 geographical races, including semi-wild and cultivated types) were selected from 3526 germplasms for de novo high-quality genome assembly. A graph pan-genome was built to analyze genomic structures and evolutionary trajectories, and PAV-GWAS was applied to dissect the genetic basis of key agronomic traits.

3. Results

• Individual Genome Assembly

High-integrity and high-continuity genomes of 107 upland cotton materials were de novo assembled and annotated, with overall quality higher than existing upland cotton genome versions (Figure a). The key quality indicators: LTR assembly index > 22, average quality value ~70, assembly quality indicator > 99.6%, with genome size 2.3-2.36 Gb, contig N50 60.97-108.61 Mb, Hi-C anchoring rate > 99%, and BUSCO completeness ~ 99.5% for gene annotation.

• Pan-Genome Construction

The pan-genome of core wild-cultivated germplasm approached saturation with 70 genomes, capturing 99% pan-gene families (Figure c). Over 51% of gene families were dispensable or specific (19% of total genes), including 1364 semi-wild-specific families enriched in defense-related phytohormone pathways.

• Cotton Evolution Analysis

The first genome-wide inversion map of upland cotton was generated, identifying 5 semi-wild-specific inversions post-divergence from Gossypium tomentosum (AD3). Materials were classified into 32 haplotypes, with modern cultivars derived from only one (narrow genetic basis). Genomic evidence verified the Central America-Caribbean dual diversity center hypothesis and reconstructed the three-stage domestication path: Yucatan Peninsula origin → Guatemala diffusion → global spread.

• PAV-GWAS Analysis

PAV-GWAS identified 69 fiber quality-related loci (62 new, undetected by SNPs) and a pleiotropic gene regulating fiber strength and seed size. It also verified the major Verticillium wilt resistance locus VW_A10 (A10 NLR cluster) and identified a new locus VW_D11 (D11).

4. Significance

This research provides a new perspective for the study of upland cotton domestication and offers important genetic markers for accelerating the process of cotton biobreeding. It not only promotes basic research on cotton but also provides technical support for the sustainable development of the cotton industry.