School of Computer and Artificial Intelligence, Zhengzhou University, Zhengzhou, 450001, Henan, China.
Yangtze Delta Region Institute (Quzhou), University of Electronic Science and Technology of China, Quzhou, China.
BMC Biol. 2024 Sep 27;22(1):216. doi: 10.1186/s12915-024-02012-x.
Drug-target interaction (DTI) prediction plays a pivotal role in drug discovery and drug repositioning, enabling the identification of potential drug candidates. However, most previous approaches often do not fully utilize the complementary relationships among multiple biological networks, which limits their ability to learn more consistent representations. Additionally, the selection strategy of negative samples significantly affects the performance of contrastive learning methods.
In this study, we propose CCL-ASPS, a novel deep learning model that incorporates Collaborative Contrastive Learning (CCL) and Adaptive Self-Paced Sampling strategy (ASPS) for drug-target interaction prediction. CCL-ASPS leverages multiple networks to learn the fused embeddings of drugs and targets, ensuring their consistent representations from individual networks. Furthermore, ASPS dynamically selects more informative negative sample pairs for contrastive learning. Experiment results on the established dataset demonstrate that CCL-ASPS achieves significant improvements compared to current state-of-the-art methods. Moreover, ablation experiments confirm the contributions of the proposed CCL and ASPS strategies.
By integrating Collaborative Contrastive Learning and Adaptive Self-Paced Sampling, the proposed CCL-ASPS effectively addresses the limitations of previous methods. This study demonstrates that CCL-ASPS achieves notable improvements in DTI predictive performance compared to current state-of-the-art approaches. The case study and cold start experiments further illustrate the capability of CCL-ASPS to effectively predict previously unknown DTI, potentially facilitating the identification of new drug-target interactions.
药物-靶点相互作用（DTI）预测在药物发现和药物重定位中起着至关重要的作用，能够识别潜在的药物候选物。然而，大多数先前的方法往往不能充分利用多个生物网络之间的互补关系，限制了它们学习更一致表示的能力。此外，负样本的选择策略对对比学习方法的性能有很大影响。
在这项研究中，我们提出了 CCL-ASPS，这是一种新的深度学习模型，它将协同对比学习（CCL）和自适应自定步采样策略（ASPS）结合起来用于药物-靶点相互作用预测。CCL-ASPS 利用多个网络来学习药物和靶点的融合嵌入，确保它们从单个网络中具有一致的表示。此外，ASPS 动态选择更具信息量的负样本对进行对比学习。在已建立的数据集上的实验结果表明，CCL-ASPS 与当前最先进的方法相比取得了显著的改进。此外，消融实验证实了所提出的 CCL 和 ASPS 策略的贡献。
通过整合协同对比学习和自适应自定步采样，所提出的 CCL-ASPS 有效地解决了先前方法的局限性。本研究表明，与当前最先进的方法相比，CCL-ASPS 在 DTI 预测性能方面取得了显著的提高。案例研究和冷启动实验进一步说明了 CCL-ASPS 有效预测先前未知的 DTI 的能力，可能有助于识别新的药物-靶点相互作用。
