Modeling and insights into the structural characteristics of drug-induced autoimmune diseases

Abstract

The incidence and complexity of drug-induced autoimmune diseases (DIAD) have been on the rise in recent years, which may lead to serious or fatal consequences. Besides, many environmental and industrial chemicals can also cause DIAD. However, there are few effective approaches to estimate the DIAD potential of drugs and other chemicals currently, and the structural characteristics and mechanism of action of DIAD compounds have not been clarified. In this study, we developed the in silico models for chemical DIAD prediction and investigated the structural characteristics of DIAD chemicals based on the reliable drug data on human autoimmune diseases. We collected 148 medications which were reported can cause DIAD clinically and 450 medications that clearly do not cause DIAD. Several different machine learning algorithms and molecular fingerprints were combined to develop the in silico models. The best performed model provided the good overall accuracy on validation set with 76.26%. The model was made freely available on the website http://diad.sapredictor.cn/. To further investigate the differences in structural characteristics between DIAD chemicals and non-DIAD chemicals, several key physicochemical properties were analyzed. The results showed that AlogP, molecular polar surface area (MPSA), and the number of hydrogen bond donors (nHDon) were significantly different between the DIAD and non-DIAD structures. They may be related to the DIAD toxicity of chemicals. In addition, 14 structural alerts (SA) for DIAD toxicity were detected from predefined substructures. The SAs may be helpful to explain the mechanism of action of drug induced autoimmune disease, and can used to identify the chemicals with potential DIAD toxicity. The structural alerts have been integrated in a structural alert-based web server SApredictor (http://www.sapredictor.cn). We hope the results could provide useful information for the recognition of DIAD chemicals and the insights of structural characteristics for chemical DIAD toxicity.

A machine learning approach to predict drug-induced autoimmunity using transcriptional data

“…There is more that could be done to refine our approach and improve the accuracy of its prediction. For example, as noted previously, other groups have had success with using structural data to predict DIA (Wu et al, 2021;Guo et al, 2022). It may therefore be possible to combine transcriptional with structural data.…”

“…One such model demonstrated that drugs containing benzene with a nitrogen containing substituent were significantly associated with a higher risk of inducing autoimmunity (Wu et al, 2021). Others have compared different machine learning methods and different molecular fingerprint packages to predict DIA using molecular properties and structural alerts (Guo et al, 2022). Although helpful, chemical structure alone may not be sufficient to determine the risk of DIA in patients, and whilst structural analysis of chemical compounds is highly sensitive, it has the potential for generating false positives (Stepan et al, 2011).…”

Modeling and Insights into the Structural Characteristics of Chemical Mitochondrial Toxicity

“…In recent years, mitochondria-related research has been concerned with drug discovery and ecotoxicology. , In vivo tests are often difficult to detect mitochondrial toxicity due to poor responses in laboratory animals, and they always manifest as organ toxicity, which cannot be easily determined as mitochondrial toxicity. In vitro studies seem more suitable, but they also require specialized equipment and cost a lot. , In silico methods, especially quantitative structure–activity relationship (QSAR) with machine learning and deep learning tools, can be used to develop artificial intelligence (AI) models by synthesizing data from in vivo and in vitro experiments to better understand and predict the chemical mitochondrial toxicity. − There were several different in silico models developed for the assessment of mitochondrial toxicity of pharmaceuticals and environmental chemicals . Zhang et al developed the support vector machine (SVM) model of mitochondrial toxicity using a dataset containing 288 small molecular drugs.…”

Machine Learning Modeling and Insights into the Structural Characteristics of Drug-Induced Neurotoxicity

“…16,17 The computational toxicology tools can be used to prioritize chemical tests and further provide some mechanistic insights. 18,19 It is of great interest to develop computational tools to assess chemical neurotoxicity potential. The machine learning modeling was widely used to estimate the toxicity potential of chemical substances.…”

“…Computational toxicology is rapidly evolving with the proliferation of commercial databases and advances in modern computational tools for the analysis of large data sets, − since it can integrate a variety of information and data to complement and extend traditional experimental toxicology research methods and encompasses multiple disciplines. , The computational toxicology tools can be used to prioritize chemical tests and further provide some mechanistic insights. , …”