Multi-Task Cascade Forest Framework for Predicting Acute Toxicity across Species

Evaluating chemical toxicity and its potential hazards to human health and the environment is essential in diverse fields, including medicine, industry, and agriculture. Multi-species acute toxicity prediction (MSATP) is critical in toxicity assessment. Traditional methods rely on exposing animals to a single high dose of a compound and observing its toxicity. However, with growing ethical concerns regarding animal testing, advancements in computational technology have positioned the artificial intelligence-based MSATP as an efficient alternative. Current research on MSATP commonly employs multi-task deep neural networks for modeling. However, the small size, high dimensionality, and sparsity of MSATP tabular data render them unsuitable for neural network approaches. To address this, we proposed a multi-task cascade forest framework for MSATP. This framework integrated feature enhancement through knowledge transfer, sample enhancement using a greedy search strategy with the covariance distance measure. The framework accommodated tasks of varying sizes in multi-task learning and was specifically designed for tabular data, achieving a 12% improvement in performance compared to current state-of-the-art methods (R ² equals to 0.64, root mean square error equals to 0.57). In a single-view context, we conducted ablation experiments to validate the effectiveness of the data enhancement strategy and introduced external dataset experiments to assess the generalization capability of the proposed method for cross-species prediction. In a multi-view context, the feature fusion method and consensus ensemble were demonstrated to further enhance the model performance. Additionally, we analyzed feature importance vectors to provide interpretable insights into species toxicity correlations. Overall, this framework effectively addressed MSATP tasks and exhibited substantial potential for application in various toxicity prediction domains.