Search Results (2)

The accumulation of senescent cells is a major contributor to aging and various age-related diseases, making developing senolytic compounds that are capable of clearing these cells an important area of research. However, progress has been hampered by the limited number of known senolytics and the incomplete understanding of their mechanisms. This study presents a powerful senolytic predictor built using phenotypic data and machine learning techniques to identify compounds with potential senolytic activity. A comprehensive training dataset consisting of 111 positive and 3951 negative compounds was curated from the literature. The dataset was used to train machine learning models, incorporating traditional molecular fingerprints, molecular descriptors, and MoLFormer molecular embeddings. By applying MoLFormer-based oversampling and testing different algorithms, it was found that the Support Vector Machine (SVM) and Multilayer Perceptron (MLP) models with MoLFormer embeddings exhibited the best performance, achieving Area Under the Curve (AUC) scores of 0.998 and 0.997, and F1 scores of 0.948 and 0.941, respectively. This senolytic predictor was then used to perform virtual screening of compounds from the DrugBank and TCMbank databases. In the DrugBank database, 98 structurally novel candidate compounds with potential senolytic activity were identified. For TCMbank, 714 potential senolytic compounds were predicted and 81 medicinal herbs with possible senolytic properties were identified. Moreover, pathway enrichment analysis revealed key targets and potential mechanisms underlying senolytic activity. In an experimental screening of predicted compounds, panaxatriol was found to exhibit senolytic activity on the etoposide-induced senescence of the IMR-90 cell line. Additionally, voclosporin was found to extend the lifespan of C. elegans more effectively than metformin, demonstrating the value of our model for drug repurposing. This study not only provides an efficient framework for discovering novel senolytic agents, but also highlights the predicted novel senolytic compounds and herbs as valuable starting points for future research into senolytic drug development. Full article

The accurate prediction of adverse drug reactions (ADRs) is essential for comprehensive drug safety evaluation. Pre-trained deep chemical language models have emerged as powerful tools capable of automatically learning molecular structural features from large-scale datasets, showing promising capabilities for the downstream prediction of molecular properties. However, the performance of pre-trained chemical language models in predicting ADRs, especially idiosyncratic ADRs induced by marketed drugs, remains largely unexplored. In this study, we propose MoLFormer-XL, a pre-trained model for encoding molecular features from canonical SMILES, in conjunction with a CNN-based model to predict drug-induced QT interval prolongation (DIQT), drug-induced teratogenicity (DIT), and drug-induced rhabdomyolysis (DIR). Our results demonstrate that the proposed model outperforms conventional models applied in previous studies for predicting DIQT, DIT, and DIR. Notably, an analysis of the learned linear attention maps highlights amines, alcohol, ethers, and aromatic halogen compounds as strongly associated with the three types of ADRs. These findings hold promise for enhancing drug discovery pipelines and reducing the drug attrition rate due to safety concerns. Full article