QSAR Study for Carcinogenic Potency of Aromatic Amines Based on GEP and MLPs

QSAR and molecular docking studies on designing potent inhibitors of SARS-CoVs main protease

Background: Severe acute respiratory syndrome coronavirus (SARS-CoVs) have emerged as a global health threat, which had caused a high rate of mortality. There is an urgent need to find effective drugs against these viruses. Objective: This study aims to predict the activity of unsymmetrical aromatic disulfides by constructing a QSAR model, and to design new compounds according to the structural and physicochemical attributes responsible for higher activity towards SARS-CoVs main protease. Methods: All molecules were constructed in ChemOffice software and molecular descriptors were calculated by CODESSA software. A regression-based linear heuristic method was established by changing descriptors datasets and calculating predicted IC₅₀ values of compounds. Then, some new compounds were designed according to molecular descriptors from the heuristic method model. The compounds with predicted values smaller than a set point were constantly screened out. Finally, the properties analysis and molecular docking were conducted to further understand the structure-activity relationships of these finalized compounds. Results: The heuristic method explored the various descriptors responsible for bioactivity and gained the best linear model with R² 0.87. The success of the model fully passed the testing set validation, proving that the model has both high statistical significance and excellent predictive ability. A total of 5 compounds with ideal predicted IC₅₀ were found from the 96 newly designed derivatives and their properties analyze was carried out. Molecular docking experiments were conducted for the optimal compound 31a, which has the best compound activity with good target protein binding capability. Conclusion: The heuristic method was quite reliable for predicting IC₅₀ values of unsymmetrical aromatic disulfides. The present research provides meaningful guidance for further exploration of the highly active inhibitors for SARS-CoVs.

Data-Driven Quantitative Structure-Activity Relationship Modeling for Human Carcinogenicity by Chronic Oral Exposure

Traditional methodologies for assessing chemical toxicity are expensive and time-consuming. Computational modeling approaches have emerged as low-cost alternatives, especially those used to develop quantitative structure-activity relationship (QSAR) models. However, conventional QSAR models have limited training data, leading to low predictivity for new compounds. We developed a data-driven modeling approach for constructing carcinogenicity-related models and used these models to identify potential new human carcinogens. To this goal, we used a probe carcinogen dataset from the US Environmental Protection Agency's Integrated Risk Information System (IRIS) to identify relevant PubChem bioassays. Responses of 25 PubChem assays were significantly relevant to carcinogenicity. Eight assays inferred carcinogenicity predictivity and were selected for QSAR model training. Using 5 machine learning algorithms and 3 types of chemical fingerprints, 15 QSAR models were developed for each PubChem assay dataset. These models showed acceptable predictivity during 5-fold cross-validation (average CCR = 0.71). Using our QSAR models, we can correctly predict and rank 342 IRIS compounds' carcinogenic potentials (PPV = 0.72). The models predicted potential new carcinogens, which were validated by a literature search. This study portends an automated technique that can be applied to prioritize potential toxicants using validated QSAR models based on extensive training sets from public data resources.

Application of a developed triple-classification machine learning model for carcinogenic prediction of hazardous organic chemicals to the US, EU, and WHO based on Chinese database

Cancer, the second largest human disease, has become a major public health problem. The prediction of chemicals' carcinogenicity before their synthesis is crucial. In this paper, seven machine learning algorithms (i.e., Random Forest (RF), Logistic Regression (LR), Support Vector Machines (SVM), Complement Naive Bayes (CNB), K-Nearest Neighbor (KNN), XGBoost, and Multilayer Perceptron (MLP)) were used to construct the carcinogenicity triple classification prediction (TCP) model (i.e., 1A, 1B, Category 2). A total of 1444 descriptors of 118 hazardous organic chemicals were calculated by Discovery Studio 2020, Sybyl X-2.0 and PaDEL-Descriptor software. The constructed carcinogenicity TCP model was evaluated through five model evaluation indicators (i.e., Accuracy, Precision, Recall, F1 Score and AUC). The model evaluation results show that Accuracy, Precision, Recall, F1 Score and AUC evaluation indicators meet requirements (greater than 0.6). The accuracy of RF, LR, XGBoost, and MLP models for predicting carcinogenicity of Category 2 is 91.67%, 79.17%, 100%, and 100%, respectively. In addition, the constructed machine learning model in this study has potential for error correction. Taking XGBoost model as an example, the predicted carcinogenicity level of 1,2,3-Trichloropropane (96-18-4) is Category 2, but the actual carcinogenicity level is 1B. But the difference between Category 2 and 1B is only 0.004, indicating that the XGBoost is one optimum model of the seven constructed machine learning models. Besides, results showed that functional groups like chlorine and benzene ring might influence the prediction of carcinogenic classification. Therefore, considering functional group characteristics of chemicals before constructing the carcinogenicity prediction model of organic chemicals is recommended. The predicted carcinogenicity of the organic chemicals using the optimum machine leaning model (i.e., XGBoost) was also evaluated and verified by the toxicokinetics. The RF and XGBoost TCP models constructed in this paper can be used for carcinogenicity detection before synthesizing new organic substances. It also provides technical support for the subsequent management of organic chemicals.

Bis-isonicotinoyl linkers containing polyaromatic scaffolds: synthesis, structure and spectroscopic properties

In this study, a new series of extended linkers containing different polyaromatic chromophores (biphenyl, naphthalene, anthracene, fluorene, 9,9-dimethylfluorene and fluorenone) functionalized with isonicotinoyl moieties have been synthesized by Pd-catalyzed cross-coupling reactions involving isonicotinamide and the appropriate aromatic dibromide. The optimized protocol led to the isolation of the target molecules in good yield and with high purity. These were characterized by ¹H NMR, FTIR, MS, and elemental analysis and their solid state structures were solved by single-crystal X-ray diffraction analysis. Electronic absorption and emission spectra were collected both in solution (DMF) and in the solid state. TDDFT calculations were carried out to investigate the effect of the isonicotinoyl moieties on the spectral features of the central chromophores. Although in solution only the linker containing a fluorenone scaffold shows a weak fluorescence, all the isolated linkers turned out to be fluorescent in the solid state, thus paving the way for their use for the fabrication of fluorescent MOFs.

Studies on the IC50 of Metabolically Stable 1-(3,3-diphenylpropyl)- piperidinyl Amides and Ureas as Human CCR5 Receptor Antagonists Based on QSAR

Background:

In this study, modulators of human Chemotactic cytokine receptor 5 (CCR5) were described using a quantitative structure-activity relationship model (QSAR). This model was based on the molecule’s chemical structure.

Methods::

All 56 compounds of CCR5 receptor antagonists were randomly separated into two sets, 43 were reserved for training and the other 13 for testing. In the course of this study, molecular models were drawn using ChemDraw software. By means of Hyperchem software as well as optimized with both AM1 (semi-empirical self-consistent-field molecular orbital) and MM+ (molecular mechanics plus force field), molecular models were described through numerous descriptors using CODESSA software.

Results:

Linear models were obtained by Heuristic Method (HM) software and nonlinear models were obtained using APS software with optimal descriptor combinations used to build linear QSAR models, involving a group of selected descriptors. As a result, values of the above two different sets were shown to result from 0.82 in testing and 0.86 in training in HM while 0.83 in testing and 0.88 in training in Gene Expression Programming (GEP).

Conclusion:

From this method, the activity of molecules could be predicted, and the molecular structure could be changed to alter its IC50, avoiding the testing of large numbers of compounds.

Virtual Screening of Anti-Cancer Compounds: Application of Monte Carlo Technique

Possibility and necessity of standardization of predictive models for anti-cancer activity are discussed. The hypothesis about rationality of common quantitative analysis of anti-cancer activity and carcinogenicity is developed. Potential of optimal descriptors to be used as a tool to build up predictive models for anti-cancer activity is examined from practical point of view. Various perspectives of application of optimal descriptors are reviewed. Stochastic nature of phenomena which are related to carcinogenic potential of various substances can be successfully detected and interpreted by the Monte Carlo technique. Hypothesises related to practical strategy and tactics of the searching for new anticancer agents are suggested.