{
  "id": 15105501,
  "title": "CucurbitAgent: An AI agent for exploring cucurbit multi-omics resources",
  "url": "https://urgent.news/2026/10/16/cucurbitagent-an-ai-agent-for-exploring-cucurbit-multi-omics-resources",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-16T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.10.09.758138v1?rss=1"
  },
  "original_language": "en",
  "account": "CucurbitAgent is an innovative AI research tool designed to streamline functional genomics studies in the Cucurbitaceae family, a group of horticultural crops of significant economic importance. This platform integrates multi-omics data from nine cultivated cucurbit species and over 18,000 scientific publications, covering approximately 256,000 genes, tissue-specific expression profiles, more than 16 million genetic variants, and roughly 242,000 predicted protein structures, along with nearly 14 million annotated binding-site residues.\n\nCucurbitAgent's primary strength lies in its ability to provide answers that are directly traceable to source records, such as gene entries, expression measurements, structural predictions, and relevant literature passages. This feature significantly enhances the reliability and credibility of the information provided by the AI.\n\nWhen evaluated against nine general-purpose large language models (LLMs) in a blind test involving 50 tasks, CucurbitAgent demonstrated an impressive 96.0% win rate. The evaluation considered factors such as factual accuracy, comprehensiveness, relevance, and traceability of the responses.\n\nThe platform is designed to be user-friendly and accessible, with its interface supporting 37 languages spoken by an estimated 70% or more of the global population. This broad language support ensures that researchers worldwide can benefit from CucurbitAgent's capabilities.\n\nBeyond simply retrieving and interpreting information from various datasets, CucurbitAgent also excels at uncovering hidden associations across different datasets. By doing so, it aids researchers in formulating testable hypotheses, thereby accelerating the process of identifying candidate genes for further investigation.\n\nThe framework of CucurbitAgent is not limited to the Cucurbitaceae family. Its design is flexible and adaptable, making it a potential blueprint for creating specialized AI agents in other domains. This versatility opens up new possibilities for AI-assisted research across a wide range of scientific fields.",
  "summary": "Functional genomics in crops increasingly depends on integrating information from an extensive literature and heterogeneous multi-omics datasets. Large language models (LLMs) offer a powerful approach to synthesizing these resources and revealing hidden connections across datasets; however, they frequently hallucinate, producing plausible but unsupported claims, when responding to gene-specific…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}