ChemBioSim: Enhancing Conformal Prediction of In Vivo Toxicity by Use of Predicted Bioactivities

Recognizing high-priority disinfection byproducts based on experimental and predicted endocrine disrupting data: Virtual screening and in vitro study

So far, about 130 disinfection by-products (DBPs) and several DBPs-groups have had their potential endocrine-disrupting effects tested on some endocrine endpoints. However, it is still not clear which specific DBPs, DBPs-groups/subgroups may be the most toxic substances or groups/subgroups for any given endocrine endpoint. In this study, we attempt to address this issue. First, a list of relevant DBPs was updated, and 1187 DBPs belonging to 4 main-groups (aliphatic, aromatic, alicyclic, heterocyclic) and 84 subgroups were described. Then, the high-priority endocrine endpoints, DBPs-groups/subgroups, and specific DBPs were determined from 18 endpoints, 4 main-groups, 84 subgroups, and 1187 specific DBPs by a virtual-screening method. The results demonstrate that most of DBPs could not disturb the endocrine endpoints in question because the proportion of active compounds associated with the endocrine endpoints ranged from 0 (human thyroid receptor beta) to 32% (human transthyretin (hTTR)). All the endpoints with a proportion of active compounds greater than 10% belonged to the thyroid system, highlighting that the potential disrupting effects of DBPs on the thyroid system should be given more attention. The aromatic and alicyclic DBPs may have higher priority than that of aliphatic and heterocyclic DBPs by considering the activity rate and potential for disrupting effects. There were 2 (halophenols and estrogen DBPs), 12, and 24 subgroups that belonged to high, moderate, and low priority classes, respectively. For individual DBPs, there were 23 (2%), 193 (16%), and 971 (82%) DBPs belonging to the high, moderate, and low priority groups, respectively. Lastly, the hTTR binding affinity of 4 DBPs was determined by an in vitro assay and all the tested DBPs exhibited dose-dependent binding potency with hTTR, which was consistent with the predicted result. Thus, more efforts should be performed to reveal the potential endocrine disruption of those high research-priority main-groups, subgroups, and individual DBPs.

Federated Learning in Computational Toxicology: An Industrial Perspective on the Effiris Hackathon

In silico approaches have acquired a towering role in pharmaceutical research and development, allowing laboratories all around the world to design, create, and optimize novel molecular entities with unprecedented efficiency. From a toxicological perspective, computational methods have guided the choices of medicinal chemists toward compounds displaying improved safety profiles. Even if the recent advances in the field are significant, many challenges remain active in the on-target and off-target prediction fields. Machine learning methods have shown their ability to identify molecules with safety concerns. However, they strongly depend on the abundance and diversity of data used for their training. Sharing such information among pharmaceutical companies remains extremely limited due to confidentiality reasons, but in this scenario, a recent concept named "federated learning" can help overcome such concerns. Within this framework, it is possible for companies to contribute to the training of common machine learning algorithms, using, but not sharing, their proprietary data. Very recently, Lhasa Limited organized a hackathon involving several industrial partners in order to assess the performance of their federated learning platform, called "Effiris". In this paper, we share our experience as Roche in participating in such an event, evaluating the performance of the federated algorithms and comparing them with those coming from our in-house-only machine learning models. Our aim is to highlight the advantages of federated learning and its intrinsic limitations and also suggest some points for potential improvements in the method.

Lessons Learned from the Grouping of Chemicals to Assess Risks to Human Health

In analogy to the periodic system that groups elements by their similarity in structure and chemical properties, the hazard of chemicals can be assessed in groups having similar structures and similar toxicological properties. Here we review case studies of chemical grouping strategies that supported the assessment of hazard, exposure, and risk to human health. By the EU-REACH and the US-TSCA New Chemicals Program, structural similarity is commonly used as the basis for grouping, but that criterion is not always adequate and sufficient. Based on the lessons learned, we derive ten principles for grouping, including: transparency of the purpose, criteria, and boundaries of the group; adequacy of methods used to justify the group; and inclusion or exclusion of substances in the group by toxicological properties. These principles apply to initial grouping to prioritize further actions as well as to definitive grouping to generate data for risk assessment. Both can expedite effective risk management.

Computational Prediction of the Phenotypic Effect of Flavonoids on Adiponectin Biosynthesis

In silico machine learning applications for phenotype-based screening have primarily been limited due to the lack of machine-readable data related to disease phenotypes. Adiponectin, a nuclear receptor (NR)-regulated adipocytokine, is relatively downregulated in human metabolic diseases. Here, we present a machine-learning model to predict the adiponectin-secretion-promoting activity of flavonoid-associated phytochemicals (FAPs). We modeled a structure-activity relationship between the chemical similarity of FAPs and their bioactivities using a random forest-based classifier, which provided the NR activity of each FAP as a probability. To link the classifier-predicted NR activity to the phenotype, we next designed a single-cell transcriptomics-based multiple linear regression model to generate the relative adiponectin score (RAS) of FAPs. In experimental validation, estimated RAS values of FAPs isolated from Scutellaria baicalensis exhibited a significant correlation with their adiponectin-secretion-promoting activity. The combined cheminformatics and bioinformatics approach enables the computational reconstruction of phenotype-based screening systems.

TOXRIC: a comprehensive database of toxicological data and benchmarks

The toxic effects of compounds on environment, humans, and other organisms have been a major focus of many research areas, including drug discovery and ecological research. Identifying the potential toxicity in the early stage of compound/drug discovery is critical. The rapid development of computational methods for evaluating various toxicity categories has increased the need for comprehensive and system-level collection of toxicological data, associated attributes, and benchmarks. To contribute toward this goal, we proposed TOXRIC (https://toxric.bioinforai.tech/), a database with comprehensive toxicological data, standardized attribute data, practical benchmarks, informative visualization of molecular representations, and an intuitive function interface. The data stored in TOXRIC contains 113 372 compounds, 13 toxicity categories, 1474 toxicity endpoints covering in vivo/in vitro endpoints and 39 feature types, covering structural, target, transcriptome, metabolic data, and other descriptors. All the curated datasets of endpoints and features can be retrieved, downloaded and directly used as output or input to Machine Learning (ML)-based prediction models. In addition to serving as a data repository, TOXRIC also provides visualization of benchmarks and molecular representations for all endpoint datasets. Based on these results, researchers can better understand and select optimal feature types, molecular representations, and baseline algorithms for each endpoint prediction task. We believe that the rich information on compound toxicology, ML-ready datasets, benchmarks and molecular representation distribution can greatly facilitate toxicological investigations, interpretation of toxicological mechanisms, compound/drug discovery and the development of computational methods.

The Evolution of Regulatory Toxicology: Where is the Gardener?

There is a need for paradigm change in the methodology employed for toxicological testing and assessment. It could be said that this change is well on its way, through an evolutionary progress analogous to that of natural selection. Darwin's Theory of Evolution has defined the idea of evolution and descendancy since the last third of the 19th century. Increasingly, this concept of 'evolution' is being applied beyond the field of biology. This Comment article discusses the progress of toxicological testing in the context of 'evolutionary pressure' and deliberates how this process can help foster the development, implementation and acceptance of mechanistic and human-relevant methods in this field. By comparing the current regulatory landscape in toxicity testing and assessment to specific elements in Charles Darwin's evolutionary theory, we aim to better understand the needs and requirements for the future.

Studying and mitigating the effects of data drifts on ML model performance at the example of chemical toxicity data

Machine learning models are widely applied to predict molecular properties or the biological activity of small molecules on a specific protein. Models can be integrated in a conformal prediction (CP) framework which adds a calibration step to estimate the confidence of the predictions. CP models present the advantage of ensuring a predefined error rate under the assumption that test and calibration set are exchangeable. In cases where the test data have drifted away from the descriptor space of the training data, or where assay setups have changed, this assumption might not be fulfilled and the models are not guaranteed to be valid. In this study, the performance of internally valid CP models when applied to either newer time-split data or to external data was evaluated. In detail, temporal data drifts were analysed based on twelve datasets from the ChEMBL database. In addition, discrepancies between models trained on publicly-available data and applied to proprietary data for the liver toxicity and MNT in vivo endpoints were investigated. In most cases, a drastic decrease in the validity of the models was observed when applied to the time-split or external (holdout) test sets. To overcome the decrease in model validity, a strategy for updating the calibration set with data more similar to the holdout set was investigated. Updating the calibration set generally improved the validity, restoring it completely to its expected value in many cases. The restored validity is the first requisite for applying the CP models with confidence. However, the increased validity comes at the cost of a decrease in model efficiency, as more predictions are identified as inconclusive. This study presents a strategy to recalibrate CP models to mitigate the effects of data drifts. Updating the calibration sets without having to retrain the model has proven to be a useful approach to restore the validity of most models.

True Grit: A Story of Perseverance Making Two out of Three the First Non-Animal Testing Strategy (Adopted as OECD Guideline No. 497)

In the last two decades, great strides have been made in developing alternative methods to animal testing for regulatory and safety testing. In 2021, a breakthrough in regulatory testing was achieved in that the first test strategies employing non-animal test methods for skin sensitization have been accepted as OECD guideline 497, which falls under the mutual acceptance of data (MAD) by OECD member states. Achieving this goal was a story of hard work and perseverance of the many people involved. This review gives an overview of some of the many aspects and timelines this entailed—just from the perspective of one stakeholder. In the end, the true grit of all involved allowed us to achieve not only a way forward in using test strategies for skin sensitization, but also a new approach to address other complex toxicological effects without the use of animals in the future.

Predicting the Skin Sensitization Potential of Small Molecules with Machine Learning Models Trained on Biologically Meaningful Descriptors

In recent years, a number of machine learning models for the prediction of the skin sensitization potential of small organic molecules have been reported and become available. These models generally perform well within their applicability domains but, as a result of the use of molecular fingerprints and other non-intuitive descriptors, the interpretability of the existing models is limited. The aim of this work is to develop a strategy to replace the non-intuitive features by predicted outcomes of bioassays. We show that such replacement is indeed possible and that as few as ten interpretable, predicted bioactivities are sufficient to reach competitive performance. On a holdout data set of 257 compounds, the best model (“Skin Doctor CP:Bio”) obtained an efficiency of 0.82 and an MCC of 0.52 (at the significance level of 0.20). Skin Doctor CP:Bio is available free of charge for academic research. The modeling strategies explored in this work are easily transferable and could be adopted for the development of more interpretable machine learning models for the prediction of the bioactivity and toxicity of small organic compounds.