Wani Mushtaq Ahmad, Kumari Pooja, Nargotra Amit
Discovery Informatics Group, NPMC Division, CSIR-Indian Institute of Integrative Medicine, Jammu, 180001, India.
Discovery Informatics Group, NPMC Division, CSIR-Indian Institute of Integrative Medicine, Jammu, 180001, India.
Comput Biol Med. 2025 Aug;194:110448. doi: 10.1016/j.compbiomed.2025.110448. Epub 2025 Jun 4.
Sphingosine kinase 1 (SphK1) plays a pivotal role in cancer progression, metastasis, and chemotherapy resistance, making it a key target for therapeutic interventions in cancer, cardiovascular diseases, and inflammation. Machine learning models, including Decision Trees (DT), Support Vector Machines (SVM), k-nearest neighbors (KNN), Random Forest (RF), and AdaBoost (ABDT), were developed to predict potential SphK1 inhibitors using a dataset of 534 compounds. RF achieved the highest performance with an accuracy of 89.81 %, specificity of 90 %, and F1-score of 0.88, while ABDT followed closely with an accuracy of 87.94 % and recall of 88.34 %. The models were further validated on a dataset of 2,173 compounds from our in-house synthetic compound library. 187 potential SphK1 inhibitors were predicted from the in-house compound library demonstrating the robustness of the models. Molecular dynamics simulations over 100 ns revealed that compounds IS01027, IS01265, and IS00998 remained stable in the sphingosine-binding pocket of SphK1, exhibiting minimal RMSD values. MM/GBSA calculations showed that PF-543 had the lowest predicted binding energy (-101.25 kcal/mol), followed by IS01265 (-94.52 kcal/mol), IS01027 (-85.53 kcal/mol), and IS00998 (-82.57 kcal/mol), indicating their strong binding affinity to the protein. Key interactions, including hydrogen bonds, hydrophobic contacts, and π-π interactions, were identified as critical for stability and binding. All four top-ranked ligands displayed strong predicted drug-like characteristics, such as high gastrointestinal absorption, and moderate oral bioavailability, with IS00998 standing out for its excellent solubility. Our integrated approach, combining machine learning, molecular docking, and molecular dynamics simulations, highlights the potential of computational methods to accelerate the discovery of SphK1 inhibitors, offering valuable insights for therapeutic innovation.
鞘氨醇激酶1（SphK1）在癌症进展、转移和化疗耐药中起关键作用，使其成为癌症、心血管疾病和炎症治疗干预的关键靶点。利用一个包含534种化合物的数据集，开发了包括决策树（DT）、支持向量机（SVM）、k近邻（KNN）、随机森林（RF）和AdaBoost（ABDT）在内的机器学习模型，以预测潜在的SphK1抑制剂。随机森林的性能最佳，准确率为89.81%，特异性为90%，F1分数为0.88，而AdaBoost紧随其后，准确率为87.94%，召回率为88.34%。这些模型在我们内部合成化合物库中的2173种化合物数据集上进一步得到验证。从内部化合物库中预测出187种潜在的SphK1抑制剂，证明了模型的稳健性。超过100纳秒的分子动力学模拟表明，化合物IS01027、IS01265和IS00998在SphK1的鞘氨醇结合口袋中保持稳定，均方根偏差值最小。MM/GBSA计算表明，PF - 543的预测结合能最低（-101.25千卡/摩尔），其次是IS01265（-94.52千卡/摩尔）、IS01027（-85.53千卡/摩尔）和IS00998（-82.57千卡/摩尔），表明它们与该蛋白质具有很强的结合亲和力。包括氢键、疏水接触和π-π相互作用在内的关键相互作用被确定为稳定性和结合的关键因素。所有四种排名靠前的配体都显示出很强的预测类药物特性，如高胃肠道吸收和适度的口服生物利用度，其中IS00998因其出色的溶解性而尤为突出。我们将机器学习、分子对接和分子动力学模拟相结合的综合方法，突出了计算方法在加速SphK1抑制剂发现方面的潜力，为治疗创新提供了有价值的见解。
