Jingjing Wang

Predicting acute toxicity across species is essential for early-stage drug safety evaluation. While recent efforts have primarily focused on improving predictive accuracy, they often fail to address two critical issues: the substantial divergence in toxicity mechanisms among different species, and the inherent noise present in experimental data. To bridge this gap, we introduce a Probabilistic Multitask Active Learning (PMAL) framework for multi-species acute toxicity prediction. Our framework integrates two key modules: a Probabilistic Multitask Learning (PML) component which jointly models the predictive distributions of multiple toxicity endpoints from a probabilistic viewpoint, and an Uncertainty-based Active Learning (UAL) component which strategically selects the most informative compounds for experimental annotation based on predictive uncertainty. Empirical evaluations demonstrate that PMAL surpasses state-of-the-art methods and is capable of providing well-calibrated uncertainty estimates for small molecules across diverse toxicity endpoints. Beyond advancing multi-species toxicity prediction, the core design principles of PMAL offer a generalizable paradigm for learning in noisy multi-task environments.

Enzyme catalytic efficiency (k_cat / K_m) is a key parameter for identifying high-activity enzymes. Recently deep learning techniques have demonstrated the potential for fast and accurate k_cat / K_m prediction. However, three challenges remain: (i) the limited size of the available k_cat / K_m dataset hinders the development of deep learning models; (ii) the model predictions lacked reliable confidence estimates; and (iii) models lacked interpretable insights into enzyme-catalyzed reactions. To address these challenges, we proposed IECata, a k_cat / K_m prediction model that provides uncertainty estimation and interpretability. IECata collected two k_cat / K_m datasets from databases and literatures. By introducing evidential deep learning, IECata provides an uncertainty estimation for k_cat / K_m predictions. Moreover, it uses bilinear attention mechanism to focused on learning crucial local interactions to interpret the key residues and substrate atoms in enzyme-catalyzed reactions. Testing results indicate that the prediction performance of IECata exceeds that of state-of-the-art benchmark models. Case studies further highlight that the incorporation of uncertainty in screening for highly active enzymes can effectively reduce false positives, thereby improving the efficiency of experimental validation and accelerating directed enzyme evolution. To public usage of IECata, we have developed an online prediction platform: http://mathtc.nscc-tj.cn/cataai/.

Enzymatic reaction kinetics are central in analyzing enzymatic reaction mechanisms and target-enzyme optimization, and thus in biomanufacturing and other industries. The enzyme turnover number (k_cat) and Michaelis constant (K_m), key kinetic parameters for measuring enzyme catalytic efficiency, are crucial for analyzing enzymatic reaction mechanisms and the directed evolution of target enzymes. Experimental determination of k_cat and K_m is costly in terms of time, labor, and cost. To consider the intrinsic connection between k_cat and K_m and further improve the prediction performance_, we propose a universal pretrained multitask deep learning model, MPEK, to predict these parameters simultaneously while considering pH, temperature, and organismal information. Through testing on the same k_cat and K_m test datasets, MPEK demonstrated superior prediction performance over the previous models. Specifically, MPEK achieved the Pearson coefficient of 0.808 for predicting k_cat, improving ca. 14.6% and 7.6% compared to the DLKcat and UniKP models, and it achieved the Pearson coefficient of 0.777 for predicting K_m, improving ca. 34.9% and 53.3% compared to the Kroll_model and UniKP models. More importantly, MPEK was able to reveal enzyme promiscuity and was sensitive to slight changes in the mutant enzyme sequence. In addition, in three case studies, it was shown that MPEK has the potential for assisted enzyme mining and directed evolution. To facilitate in silico evaluation of enzyme catalytic efficiency, we have established a web server implementing this model, which can be accessed at http://mathtc.nscc-tj.cn/mpek.