The Bigger Fish: A Comparison of Meta-Learning QSAR Models on Low-Resourced Aquatic Toxicity Regression Tasks

Abstract

Toxicological information as needed for risk assessments of chemical compounds is often sparse. Unfortunately, gathering new toxicological information experimentally often involves animal testing. Simulated alternatives, e.g., quantitative structure–activity relationship (QSAR) models, are preferred to infer the toxicity of new compounds. Aquatic toxicity data collections consist of many related taskseach predicting the toxicity of new compounds on a given species. Since many of these tasks are inherently low-resource, i.e., involve few associated compounds, this is challenging. Meta-learning is a subfield of artificial intelligence that can lead to more accurate models by enabling the utilization of information across tasks. In our work, we benchmark various state-of-the-art meta-learning techniques for building QSAR models, focusing on knowledge sharing between species. Specifically, we employ and compare transformational machine learning, model-agnostic meta-learning, fine-tuning, and multi-task models. Our experiments show that established knowledge-sharing techniques outperform single-task approaches. We recommend the use of multi-task random forest models for aquatic toxicity modeling, which matched or exceeded the performance of other approaches and robustly produced good results in the low-resource settings we studied. This model functions on a species level, predicting toxicity for multiple species across various phyla, with flexible exposure duration and on a large chemical applicability domain.

Emerging contaminants: A One Health perspective

“…On the contrary, the stability, reliability, and predictability of QSARs would be enhanced if meta-learning big data were involved in development. 494 The integration of environmental and structural factors of ECs is likely to aggravate the uncertainty of QSARs because they are hardly accommodated with the significant correlation in one model, whereas it is of particular interest for augmentation of the QSAR applicability domain. Given the numerous limitations of QSARs, high uncertainty or application factors are applied to QSAR modeling outputs during early tiers of risk assessment.…”

The pursuit of accurate predictive models of the bioactivity of small molecules

“…In this study, knowledge shearing between species was key for the development of multitask models. 64 …”

Transferability of Machine Learning Models for Geogenic Contaminated Groundwaters

“…Therefore, it is hoped that this large scale model can estimate contaminant levels in data-scarce areas, demonstrating model transferability. Similar strategies are also in demand in other environmental science research areas, such as quantitative structure−activity relationships 18 and species distribution modeling. 19 In fact, clues suggest that model transferability may fail when dealing with multiple hydrogeological units.…”