[论文解读] Accurate Clinical Toxicity Prediction using Multi-task Deep Neural Nets and Contrastive Molecular Explanations

Bhanushee Sharma, Vijil Chenthamarakshan+5 more|arXiv (Cornell University)|Apr 13, 2022

Computational Drug Discovery Methods被引用 11

一句话总结

本文提出一种基于Morgan指纹和预训练SMILES嵌入的多任务深度神经网络，可同时预测体外、体内及临床毒性，在MoleculeNet基准上达到最先进性能。该方法通过相关正负特征（Pertinent Positive and Negative）提供对比解释，其结果与已知毒性药效团一致，提升了模型可解释性，并揭示了模型在体外/体内毒性药效团恢复方面优于临床终点的偏好。

ABSTRACT

Explainable ML for molecular toxicity prediction is a promising approach for efficient drug development and chemical safety. A predictive ML model of toxicity can reduce experimental cost and time while mitigating ethical concerns by significantly reducing animal and clinical testing. Herein, we use a deep learning framework for simultaneously modeling in vitro, in vivo, and clinical toxicity data. Two different molecular input representations are used: Morgan fingerprints and pre-training SMILES embeddings. A multi-task deep learning model accurately predicts toxicity for all endpoints, including clinical, as indicated by AUROC and balanced accuracy. In particular, SMILES embeddings as input to the multi-task model improved clinical toxicity predictions compared to existing models in MoleculeNet benchmark. Additionally, our multi-task approach is comprehensive in the sense that it is comparable to state-of-the-art approaches for specific endpoints in in vitro, in vivo and clinical platforms. Through both the multi-task model and transfer learning, we were able to indicate the minimal need of in vivo data for clinical toxicity predictions. To provide confidence and explain the model's predictions, we adapt a post-hoc contrastive explanation method that returns pertinent positive and pertinent negative features, which correspond well to known mutagenic and reactive toxicophores, such as unsubstituted bonded heteroatoms, aromatic amines, and Michael receptors. Furthermore, toxicophore recovery by pertinent feature analysis captures more of the in vitro (53%) and in vivo (56%), rather than of the clinical (8%), endpoints, and indeed uncovers a preference in known toxicophore data towards in vitro and in vivo experimental data. To our knowledge, this is the first contrastive explanation, using both present and absent substructures, for predictions of clinical and in vivo molecular toxicity.

研究动机与目标

• 为应对因毒性导致药物候选物高失败率的挑战，利用机器学习准确预测临床毒性。
• 开发统一的深度学习框架，同时建模体外、体内及临床毒性终点，克服单任务模型的局限性。
• 通过对比解释方法，识别分子结构中存在（Pertinent Positive）与缺失（Pertinent Negative）的关键子结构，提供可操作且忠实的毒性预测解释。
• 通过迁移学习与多任务训练，减少对昂贵且具有伦理争议的体内数据的依赖。
• 验证模型的解释结果与已知致突变性和反应性毒性药效团一致，从而增强在药物开发中的可信度与实用性。

提出的方法

• 在三个毒性终点（Tox21：体外；RTECS：体内；ClinTox：临床）上训练一个多任务深度神经网络（MT-DNN），采用共享层与任务特定层相结合的结构。
• 采用两种分子输入表示：Morgan指纹与预训练SMILES嵌入，后者在临床毒性预测中表现更优。
• 通过微调策略应用迁移学习：在Tox21与RTECS上预训练后，仅用1个周期（epoch）在ClinTox任务上微调模型。
• 将对比解释方法（CEM）适配用于生成Pertinent Positive（PP）与Pertinent Negative（PN）特征，通过带L1与L2正则化的优化问题求解，并引入基于自编码器的重构损失。
• 优化过程采用投影快速迭代收缩阈值算法（FISTA），以寻找使模型预测发生改变的最小且必要的输入分子扰动。
• 使用自编码器确保扰动后的分子保持化学合理性，通过γ加权正则化强制执行重构损失。

实验结果

研究问题

• RQ1 多任务深度学习模型是否能同时且准确地使用共享与任务特定表示，预测体外、体内及临床平台的毒性？
• RQ2 与传统分子指纹相比，使用预训练SMILES嵌入是否能提升临床毒性预测性能？
• RQ3 迁移学习在多大程度上可减少临床毒性预测中对体内数据的依赖？
• RQ4 对比解释（Pertinent Positives与Negatives）能否识别出已知毒性药效团，如未取代杂原子、芳香胺及迈克尔受体？
• RQ5 为何模型的解释性能与体外和体内毒性药效团的相关性更强，而与临床毒性药效团的相关性较弱？

主要发现

• 采用预训练SMILES嵌入的多任务模型在MoleculeNet基准上实现了临床毒性预测的最先进性能，优于现有模型。
• 该模型在所有平台上均表现出高预测准确性，曲线下面积（AUC）与平衡准确率指标表明其具有强大的泛化能力。
• 通过Pertinent Positive与Negative特征识别毒性药效团，成功捕获了53%的已知体外毒性药效团与56%的体内毒性药效团，但仅捕获8%的临床毒性药效团。
• 毒性药效团识别率的差异表明，已知毒性药效团数据库存在偏向体外与体内实验数据的数据偏差。
• 对比解释成功识别出芳香胺、未取代杂原子及迈克尔受体等生物相关亚结构，确认其为毒性预测的关键贡献因素。
• 迁移学习的应用使模型在极少体内数据条件下仍能实现有效的临床毒性预测，证明了其数据效率及减少动物实验的潜力。