DrugReasoner:可解释的药物批准预测,使用推理增强型语言模型
Mohammadreza Ghaffarzadeh-Esfahani1, Ali Motahharynia1,2, Nahid Yousefian1
1Regenerative Medicine Research Center, Isfahan University of Medical Sciences, Isfahan, Iran.
PloS one
|February 19, 2026
概括
DrugReasoner是一种新型的大型语言模型 (LLM),可以高精度地预测小分子药物批准,并提供明确的理由. 这种可解释的AI工具增强了药物发现中的决策.
科学领域:
- 计算化学是一种计算化学.
- 人工智能在药物发现中的作用
- 机器学习用于制药开发.
背景情况:
- 药物发现是资源密集型的,需要准确的早期预测批准结果.
- 经典和深度学习模型提供预测能力,但缺乏可解释性.
- 可解释性对于优化研究投资和增强对人工智能驱动决策的信任至关重要.
研究的目的:
- 开发和评估DrugReasoner,一个基于推理的大型语言模型 (LLM),用于预测小分子药物批准.
- 通过提供逐步推理,提高AI模型在药物发现中的可解释性.
- 在制药决策中提高人工智能驱动预测的准确性和稳定性.
主要方法:
- 使用LLaMA架构开发了DrugReasoner,并通过组相对策略优化 (GRPO) 进行了微调.
- 集成的分子描述器与相似的批准/未批准化合物进行比较推理.
- 使用验证,测试和外部独立数据集的AUC和F1分数来评估性能.
主要成果:
- DrugReasoner以AUC为0.732 (验证) 和0.725 (测试) 取得了强的表现,表现优于传统基线.
- 在一个独立的数据集上,DrugReasoner实现了0.728的AUC和0.774的F1得分,超过了ChemAP模型.
- 该模型提供了可解释的,逐步推理,以及预测和信心评分.
结论:
- 像DrugReasoner这样的推理增强的LLM提供了一种有希望的方法,用于可解释和准确的AI辅助药物发现.
- 通过将预测准确性与透明度相结合,DrugReasoner解决了人工智能驱动的制药决策中的一个关键瓶.
- 该模型的强大性能和可解释性证明了其在优化药物开发管道的现实应用潜力.
更多相关视频
03:14Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
1.3K
05:47Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
Published on: June 13, 2025
1.9K
相关概念视频
Inductive Reasoning
59.0K
Inductive reasoning is a form of logical thinking that uses related observations to arrive at a general conclusion. It is uncertain and operates in degrees to which the conclusions are credible. As such, inductive arguments can be weak or strong, rather than valid or invalid, and conclusions can be used to formulate testable, falsifiable hypotheses.
Inductive reasoning is common in descriptive science. A life scientist makes observations and records them. This data can be qualitative or...
Inductive reasoning is common in descriptive science. A life scientist makes observations and records them. This data can be qualitative or...
59.0K
Deductive Reasoning
59.5K
Deductive reasoning, or deduction, is the type of logic used in hypothesis-based science. In deductive reasoning, the pattern of thinking moves in the opposite direction as compared to inductive reasoning, which means that it uses a general principle or law to predict specific results. From those general principles, a scientist can deduce and predict the specific results that would be valid as long as the general principles are valid.
For example, a researcher can deduce specific predictions...
For example, a researcher can deduce specific predictions...
59.5K
Drug Regulation
3.0K
Drug regulation encompasses the management of drug usage by evaluating its safety and efficacy through assessments conducted by regulatory authorities. Regrettably, the history of drug regulation is marred by several catastrophic events. One such incident is the Elixir Sulfanilamide tragedy, in which the toxic compound diethyl glycol was included in a sweet-tasting medication, leading to numerous fatalities. This event prompted the enactment of the Food, Drug, and Cosmetic Act in 1938. Under...
3.0K
Pharmacovigilance
2.0K
Post-marketing surveillance is a critical component of pharmaceutical regulation, often uncovering unanticipated adverse drug reactions (ADRs) once a drug is widely used over an extended period.
This process, termed pharmacovigilance, aims to detect, evaluate, and minimize harmful effects related to medication use. The data collection for pharmacovigilance depends on spontaneous reporting systems, where healthcare professionals or patients voluntarily report suspected ADRs.
In some cases, there...
This process, termed pharmacovigilance, aims to detect, evaluate, and minimize harmful effects related to medication use. The data collection for pharmacovigilance depends on spontaneous reporting systems, where healthcare professionals or patients voluntarily report suspected ADRs.
In some cases, there...
2.0K
Impact of Pharmacokinetic–Pharmacodynamic Models: Regulatory Decisions
97
PK–PD modeling has significantly influenced FDA regulatory decisions, particularly drug approval, dosage optimization, and labeling. These models integrate pharmacokinetics (PK) and pharmacodynamics (PD) to predict drug behavior and effects, aiding in optimizing dosing regimens and enhancing the probability of clinical trial success.One notable example is Nesiritide (Natrecor®), a recombinant human brain natriuretic peptide for treating acute decompensated congestive heart failure...
97
Drug Toxicity: Risk factors
248
Adverse Drug Reactions (ADRs) are potential complications that arise during pharmacotherapy, influenced by multiple risk factors. Age plays a significant role; both neonates and the elderly are at heightened risk due to their respective immature and diminished metabolic and elimination processes. Gender also impacts ADRs, with females experiencing a 1.5 to 1.7-fold greater risk than males, which may be linked to pharmacokinetic, pharmacodynamic, and hormonal differences. Notably, neonates, the...
248
