Integrating image-based phenotyping and GWAS to map resistance to spittlebug nymphs in interspecific Urochloa grasses

Abstract

Urochloa grasses are among the most widely used forage grasses across the tropics. Spittlebugs (Hemiptera: Cercopidae) are major pests of tropical Urochloa (syn. Brachiaria) grass pastures, severely reducing forage productivity and quality. Understanding the genetic basis of host-plant resistance is essential for developing durable resistant cultivars. Here, we combined high-throughput image-based phenotyping and genome-wide association studies (GWAS) to dissect the genetic architecture of response to Aeneolamia varia nymphs in 339 interspecific F1 hybrids derived from crosses between resistant sexual and susceptible apomictic Urochloa parents. Digital image analysis using both unsupervised (DQU) and supervised (DTR) quantification pipelines enabled accurate estimation of plant damage, yielding moderate to high broad-sense heritability estimates (H2 = 0.49 to 0.66). In contrast, insect survival (NTS) exhibited low to moderate correlations with all damage traits and lower heritability estimates (H2 = 0.42). Using 57,051 high-quality SNPs aligned to the genome of the hybrid cultivar Basilisk, GWAS models identified 18 quantitative trait loci (QTLs) for plant damage traits, but none for insect survival (antibiosis). Six robust QTLs on chromosomes 1, 6, 7, 27, 29, and 36 were consistently detected across models and phenotyping methods, explaining up to 21.5% of phenotypic variance. Candidate gene analysis revealed proteins involved in hormone signaling, oxidative stress response, and cell wall modification, suggesting multifaceted plant-insect interaction mechanisms. These results provide a foundational set of molecular markers associated with spittlebug response in Urochloa grasses, useful for marker-assisted and genomic selection in the forage breeding program.

Plant phenotyping relevance

高スループット画像表現型解析と、植物損傷を推定する2つの画像解析パイプラインが研究の中心であり、異なる手法間の比較と形質推定性能も評価している。

Code and data availability

The paper's digital plant-damage images are publicly deposited in Harvard Dataverse (paper-specific phenotyping input). The RAD-Seq accession PRJEB109285 is a sequencing/omics deposit and is excluded per criteria. No author analysis code repository with explicit availability URL is stated.