Advancing Clinical Genomic Analysis with a Skill-Centred Agent

Abstract

Clinical genomic analysis requires experts to integrate evolving guidelines, databases, and computational evidence. As sequencing has become increasingly accessible, this complex and labor-intensive process has become a major bottleneck to the timely use of genomic information in clinical decision-making. To fill this gap, we developed ClinGeX, a large-language-model agent that composes 105 curated executable skills through task-specific retrieval and dependency-aware orchestration. On 223 temporally held-out benchmark cases, ClinGeX achieved 71.7\% macro-averaged task-level concordance with expert reference classifications, compared with 45.7\% for the strongest comparator. In a three-condition, counterbalanced human-AI collaboration study, 45 clinical genomic professionals assessed 135 clinically derived cases under routine (no AI assistance), DeepSeek-assisted and ClinGeX-assisted conditions. ClinGeX assistance increased raw case-level concordance across criterion-level variant analysis, genomic interpretation in rare disease and molecular oncology decision-making (51.1\%, 57.8\%, 64.4\% versus 15.6\%, 31.1\%, 20.0\% in comparator conditions) and reduced reported interpretation time in all scenarios. During an exploratory one-month real-world non-interventional assessment, 11 experienced professionals submitted feedback on 96 self-selected analysis episodes and reported frequent conclusion adoption and time savings. These findings support executable domain skills as a promising approach to augment specialist genomic interpretation.

ddkg.skill: A Compositional Agent Skill for Translating Biomedical and Bioinformatics Questions into Cypher for the Data Distillery Knowledge Graph
D. M. Taylor, A. Lahiri, B. Stear +7
bioRxiv26 Sep 2026PreprintCC BYFull textBioinformatics & computational biologybioRxiv
From code to natural language: MErlin - a multiomics toolkit for bacterial epigenomics delivered as Claude agent skill.
I. Passeri, S. Pety, M. Giovannini +3
bioRxiv2 Sep 2026PreprintCC BY-NC-NDFull textBioinformatics & computational biologybioRxiv
Population structure and diversity of cowpea ( Vigna unguiculata [L.] Walp.): Genomic inference and preliminary distribution of breeding-relevant traits in a Niger-centred Sahelian diversity panel
Hadiara Hamadou Hamidou, Abdoul-Aziz Saïdou, Abdou Harou +2
Plant Genetic Resources15 Sep 2026Peer-reviewedCC BY 4.0Full textGenetics & genomics Plant biologyHAL
Isolation, identification, and comparative genomic analysis of Vibrio harveyi as a causative agent for a skin ulcer disease of endangered fish Chinese bahaba (Bahaba taipingensis)
L. Zhong, R. Zhu, Y. Zhang +4
bioRxiv10 Sep 2026PreprintAll rights reservedFull textMicrobiologybioRxiv
Human iPSC-Derived Brain Cells, Organoids, and Organoids-on-Chips: Advancing Neurological Disease Modeling to Improve Clinical Outcome
Gabriel Burdman, Marco Hanna, Adriana Yndart Arias +2
Cells29 Sep 2026Peer-reviewedCC BYFull textCell biologyEurope PMC

Cite

BibTeX RIS

BibTeX

@misc{zheng2026advancing,
  title = {{Advancing Clinical Genomic Analysis with a Skill-Centred Agent}},
  author = {Qiaoyu Zheng and Weike Zhao and Yanjie Fan and Ye Cao and Mengqi Jiang and Xiao Zhou and Xiaoman Zhang and Pengcheng Qiu and Yongguo Yu and Yanfeng Wang and Ya Zhang and Chaoyi Wu and Weidi Xie},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.40556},
  eprint = {2610.40556},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.40556},
}

RIS

TY  - UNPB
TI  - Advancing Clinical Genomic Analysis with a Skill-Centred Agent
AU  - Qiaoyu Zheng
AU  - Weike Zhao
AU  - Yanjie Fan
AU  - Ye Cao
AU  - Mengqi Jiang
AU  - Xiao Zhou
AU  - Xiaoman Zhang
AU  - Pengcheng Qiu
AU  - Yongguo Yu
AU  - Yanfeng Wang
AU  - Ya Zhang
AU  - Chaoyi Wu
AU  - Weidi Xie
PY  - 2026
DA  - 2026/10/29
PB  - arXiv
DO  - 10.48550/arxiv.2610.40556
UR  - https://arxiv.org/abs/2610.40556
AB  - Clinical genomic analysis requires experts to integrate evolving guidelines, databases, and computational evidence. As sequencing has become increasingly accessible, this complex and labor-intensive process has become a major bottleneck to the timely use of genomic information in clinical decision-making. To fill this gap, we developed ClinGeX, a large-language-model agent that composes 105 curated executable skills through task-specific retrieval and dependency-aware orchestration. On 223 temporally held-out benchmark cases, ClinGeX achieved 71.7\% macro-averaged task-level concordance with expert reference classifications, compared with 45.7\% for the strongest comparator. In a three-condition, counterbalanced human-AI collaboration study, 45 clinical genomic professionals assessed 135 clinically derived cases under routine (no AI assistance), DeepSeek-assisted and ClinGeX-assisted conditions. ClinGeX assistance increased raw case-level concordance across criterion-level variant analysis, genomic interpretation in rare disease and molecular oncology decision-making (51.1\%, 57.8\%, 64.4\% versus 15.6\%, 31.1\%, 20.0\% in comparator conditions) and reduced reported interpretation time in all scenarios. During an exploratory one-month real-world non-interventional assessment, 11 experienced professionals submitted feedback on 96 self-selected analysis episodes and reported frequent conclusion adoption and time savings. These findings support executable domain skills as a promising approach to augment specialist genomic interpretation.
ER  -

Abstract

Related

Cite