How important is to detect systematic error in predictions and understand statistical applicability domain of QSAR models?

Roy, Kunal; Ambure, Pravin; Aher, Rahul Balasaheb

doi:10.1016/j.chemolab.2017.01.010

Cited by 130 publications

(61 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The reliability of the predictions was confirmed by the determination of the error bias of the model . This verifies the absence of systematic error, assuring that error values lie both above and below the null residual.…”

Section: Methodsologysupporting

confidence: 58%

Conformation‐Independent QSAR Study on Human Epidermal Growth Factor Receptor‐2 (HER2) Inhibitors

et al. 2017

View full text Add to dashboard Cite

Inhibition of HER2 (human epidermal growth factor receptor 2) expression and function is required in several cancer treatments. Numerous compounds with very different molecular structures have been suggested as HER2 inhibitors. Here we perform quantitative structure‐activity relationship (QSAR) analysis on 444 of such compounds to investigate the molecular properties that may influence its efficiency. Models based on 1D and 2D flexible molecular descriptors are proposed to develop simple models based solely on constitutional and topological molecular features. A large number of non‐conformational descriptors (17974) was used to thoroughly explore the structural characteristics that influence the HER2 inhibitory activity. Three different approaches were explored using: 1) Molecular Descriptors, 2) Flexible Molecular Descriptors, and 3) Hybrid Descriptors. A QSAR model for HER2 inhibitors was successfully developed. Some properties such as electronegativity, aromatic character, and the presence of amino groups appear as molecular characteristics that may have influence in the HER2 inhibitory activity.

show abstract

Section: Methodsologysupporting

confidence: 58%

Conformation‐Independent QSAR Study on Human Epidermal Growth Factor Receptor‐2 (HER2) Inhibitors

et al. 2017

View full text Add to dashboard Cite

show abstract

“…According to the results of the prediction evaluations, our models performed well and can markedly decrease the false‐positive rate in prediction compared with other previous models. Afterward, we chose three simple but important descriptors—molecular weight (MW), lipo‐hydro partition coefficient (LogP), topological polar surface area (TPSA)—to define the applicability domain (AD) of our model and analyzed the significance of these descriptors in the mechanism of BBB penetration . Also, substructure frequency analysis and information gain (IG) method were used to identify several representative substructures of BBB+ and BBB− by using SARpy tools.…”

Section: Introductionmentioning

confidence: 99%

In Silico Prediction of Blood–Brain Barrier Permeability of Compounds by Machine Learning and Resampling Methods

et al. 2018

View full text Add to dashboard Cite

The blood-brain barrier (BBB) as a part of absorption protects the central nervous system by separating the brain tissue from the bloodstream. In recent years, BBB permeability has become a critical issue in chemical ADMET prediction, but almost all models were built using imbalanced data sets, which caused a high false-positive rate. Therefore, we tried to solve the problem of biased data sets and built a reliable classification model with 2358 compounds. Machine learning and resampling methods were used simultaneously for the refinement of models with both 2 D molecular descriptors and molecular fingerprints to represent the chemicals. Through a series of evaluation, we realized that resampling methods such as Synthetic Minority Oversampling Technique (SMOTE) and SMOTE+edited nearest neighbor could effectively solve the problem of imbalanced data sets and that MACCS fingerprint combined with support vector machine performed the best. After the final construction of a consensus model, the overall accuracy rate was increased to 0.966 for the final external data set. Also, the accuracy rate of the model for the test set was 0.919, with an excellent balanced capacity of 0.925 (sensitivity) to predict BBB-positive compounds and of 0.899 (specificity) to predict BBB-negative compounds. Compared with other BBB classification models, our models reduced the rate of false positives and were more robust in prediction of BBB-positive as well as BBB-negative compounds, which would be quite helpful in early drug discovery.

show abstract

“…The estimated predictive power of the models is evaluated on the test set by calculating the root-mean-squared error in prediction (RMSE; error-based metric) and correlation metrics (e.g., R 2 ; correlation-based metric) for the predicted against the observed values (Kuhn & Johnson, 2013;Roy, Ambure, & Aher, 2017). The values for these metrics can be calculated using the camb functions Rsquared_CV and RMSE_CV (see section 4 of Supplementary File 3).…”

Section: Remove Uninformative Molecular Descriptors Using the Functiomentioning

confidence: 99%

Elucidating Compound Mechanism of Action and Predicting Cytotoxicity Using Machine Learning Approaches, Taking Prediction Confidence into Account

Drakakis

Cortés-Ciriano

Alexander-Dann

et al. 2019

CP Chemical Biology

View full text Add to dashboard Cite

The modes of action (MoAs) of drugs frequently are unknown, because many are small molecules initially identified from phenotypic screens, giving rise to the need to elucidate their MoAs. In addition, the high attrition rate for candidate drugs in preclinical studies due to intolerable toxicity has motivated the development of computational approaches to predict drug candidate (cyto)toxicity as early as possible in the drug-discovery process. Here, we provide detailed instructions for capitalizing on bioactivity predictions to elucidate the MoAs of small molecules and infer their underlying phenotypic effects. We illustrate how these predictions can be used to infer the underlying antidepressive effects of marketed drugs. We also provide the necessary functionalities to model cytotoxicity data using single and ensemble machine-learning algorithms. Finally, we give detailed instructions on how to calculate confidence intervals for individual predictions using the conformal prediction framework. C 2019 by John Wiley & Sons, Inc.Keywords: ChEMBL r cytotoxicity r in silico bioactivity prediction r mechanism of action r polypharmacology r toxicology modeling

show abstract

How important is to detect systematic error in predictions and understand statistical applicability domain of QSAR models?

Cited by 130 publications

References 38 publications

Conformation‐Independent QSAR Study on Human Epidermal Growth Factor Receptor‐2 (HER2) Inhibitors

Conformation‐Independent QSAR Study on Human Epidermal Growth Factor Receptor‐2 (HER2) Inhibitors

In Silico Prediction of Blood–Brain Barrier Permeability of Compounds by Machine Learning and Resampling Methods

Elucidating Compound Mechanism of Action and Predicting Cytotoxicity Using Machine Learning Approaches, Taking Prediction Confidence into Account

Contact Info

Product

Resources

About