arXiv cs.AI· Pratik Dutta, Matthew B. Obusan, Max Chao, Rekha Sathian, Nimisha Papineni, Ramana V. Davuluri·· 9 天前AI 评分24
ARGUS：证据约束的智能体框架如何解读非编码区单核苷酸变异
Unlocking the Regulatory Genome by ARGUS: An Evidence-Constrained Agentic Framework for Interpreting Single Nucleotide Variants
AI 导读
ARGUS 是一个将确定性生物计算与 LLM 推理严格分离的智能体框架，用 458 个基于 DNABERT 的转录因子结合模型在假设驱动的调查循环中解读非编码区变异。
正文 · 原文
Abstract:Over 90% of disease-associated variants from genome-wide association studies fall in noncoding regulatory regions, yet their functional interpretation remains a central open problem in genomic medicine. Large language models prompted to interpret such variants routinely hallucinate transcription factor (TF) binding changes, fabricate experimental support, and assign biological significance to statistically negligible signals. We present ARGUS (Agentic Regulatory Genomics for an Uncertainty-aware Scientist), which strictly separates deterministic biological computation from LLM-mediated reasoning. ARGUS wraps 458 DNABERT-based TF binding models in a hypothesis-directed investigation loop where a planner selects evidence sources based on current uncertainty, a verifier deterministically interprets each observation, and intermediate results change the investigation path. On variant rs6983267 at the 8q24 cancer risk locus, the same planner produces four divergent trajectories for four TFs. FOXA1 is rescued in 3 steps when real ADASTRA allele-specific binding data (15 experiments, FDR = 0.030) reveals a model false negative masked by saturation. KLF6 traverses 8 steps across ADASTRA, JASPAR motif analysis, and ENCODE cCRE regulatory annotation before abstaining due to mixed indirect evidence. RAD21 abstains in 8 steps after ADASTRA returns a coverage-qualified but nonsignificant allelic test (5 experiments, FDR = 0.65), and SP1, which shares FOXA1's saturated retained prediction, abstains because no direct experimental evidence exists at this locus. All observations come from real ADASTRA, JASPAR, and ENCODE cCRE queries; none are simulated. A comparison of fixed-priority and LLM-mediated planning shows that the LLM planner reaches identical verdicts with fewer tool calls by declining evidence that cannot resolve the claim under test.
| Comments: | Accepted at the NeurIPS 2026 Workshop on Agentic AI for Biological Discovery (AgenticLS). Code: this https URL |
| Subjects: | Genomics (q-bio.GN); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.12281 [q-bio.GN] |
| (or arXiv:2610.12281v1 [q-bio.GN] for this version) | |
| https://doi.org/10.48550/arXiv.2610.12281 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Pratik Dutta [view email]
[v1]
Thu, 8 Oct 2026 16:38:32 UTC (439 KB)
来源：arXiv cs.AI · arxiv.org