Exploring different strategies for imbalanced ADME data problem: case study on Caco-2 permeability modeling

Pham-The, Hai; Casañola-Martín, Gerardo M.; Garrigues, Teresa María; Bermejo, Marival; González‐Álvarez, Isabel; Nam, Nguyen Hai; Cabrera‐Pérez, Miguel Ángel; Le-Thi-Thu, Huong

doi:10.1007/s11030-015-9649-4

Cited by 11 publications

(6 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The present study was motivated by the scarcity of reported efforts in the application of the above-mentioned methods to the SAR-based chemical classification domain. We conducted a literature survey which only identified a few studies in this domain where cost-sensitive learning [ 28 , 29 ], resampling [ 29 , 30 ], conformal prediction [ 18 ] and extreme entropy machines [ 1 , 31 ] were employed to specifically deal with data imbalance. Although predictive modeling was improved for certain datasets, a consistent performance enhancement was not observed as a result of resampling and algorithm modification.…”

Section: Introductionmentioning

confidence: 99%

Structure–activity relationship-based chemical classification of highly imbalanced Tox21 datasets

et al. 2020

View full text Add to dashboard Cite

The specificity of toxicant-target biomolecule interactions lends to the very imbalanced nature of many toxicity datasets, causing poor performance in Structure–Activity Relationship (SAR)-based chemical classification. Undersampling and oversampling are representative techniques for handling such an imbalance challenge. However, removing inactive chemical compound instances from the majority class using an undersampling technique can result in information loss, whereas increasing active toxicant instances in the minority class by interpolation tends to introduce artificial minority instances that often cross into the majority class space, giving rise to class overlapping and a higher false prediction rate. In this study, in order to improve the prediction accuracy of imbalanced learning, we employed SMOTEENN, a combination of Synthetic Minority Over-sampling Technique (SMOTE) and Edited Nearest Neighbor (ENN) algorithms, to oversample the minority class by creating synthetic samples, followed by cleaning the mislabeled instances. We chose the highly imbalanced Tox21 dataset, which consisted of 12 in vitro bioassays for > 10,000 chemicals that were distributed unevenly between binary classes. With Random Forest (RF) as the base classifier and bagging as the ensemble strategy, we applied four hybrid learning methods, i.e., RF without imbalance handling (RF), RF with Random Undersampling (RUS), RF with SMOTE (SMO), and RF with SMOTEENN (SMN). The performance of the four learning methods was compared using nine evaluation metrics, among which F1 score, Matthews correlation coefficient and Brier score provided a more consistent assessment of the overall performance across the 12 datasets. The Friedman’s aligned ranks test and the subsequent Bergmann-Hommel post hoc test showed that SMN significantly outperformed the other three methods. We also found that a strong negative correlation existed between the prediction accuracy and the imbalance ratio (IR), which is defined as the number of inactive compounds divided by the number of active compounds. SMN became less effective when IR exceeded a certain threshold (e.g., > 28). The ability to separate the few active compounds from the vast amounts of inactive ones is of great importance in computational toxicology. This work demonstrates that the performance of SAR-based, imbalanced chemical toxicity classification can be significantly improved through the use of data rebalancing.

show abstract

Section: Introductionmentioning

confidence: 99%

Structure–activity relationship-based chemical classification of highly imbalanced Tox21 datasets

et al. 2020

View full text Add to dashboard Cite

show abstract

“…The dataset consisted of 1043 compounds with human f a , ,,,,,,,− 125 with rat f a , ,,,,,− and 484 compounds with Caco-2 P app ,,,,,,,,,,− ,,,− and human f a . When multiple permeability values were found for a single compound from more than one source, the geometric mean of the values was used as the point estimate.…”

Section: Experimental Sectionmentioning

confidence: 99%

Global Analysis of Models for Predicting Human Absorption: QSAR, In Vitro, and Preclinical Models

et al. 2021

View full text Add to dashboard Cite

Models intended to predict intestinal absorption are an essential part of the drug development process. Although many models exist for capturing intestinal absorption, many questions still exist around the applicability of these models to drug types like "beyond rule of 5" (bRo5) and low absorption compounds. This presents a challenge as current models have not been rigorously tested to understand intestinal absorption. Here, we assembled a large, structurally diverse dataset of ∼1000 compounds with known in vitro, preclinical, and human permeability and/or absorption data. In silico (quantitative structure−activity relationship), in vitro (Caco-2), and in vivo (rat) models were statistically evaluated for predictive performance against this human intestinal absorption dataset. We expect this evaluation to serve as a resource for DMPK scientists and medicinal/computational chemists to increase their understanding of permeability and absorption model utility and applications for academia and industry.

show abstract

“…PBCS was applied to evaluate the absorption profile. [34,45] In addition, the overall ADMET score compliance, especially Lipinski's RO5, bioavailability score, and characteristic toxicity were taken into account as the consensus criteria of "drug-likeness" for synthesized compounds. [31,46,47]…”

Section: Physicochemical and Admet Profilingmentioning

confidence: 99%

Design, synthesis, and evaluation of novel (E)‐N'‐(3‐allyl‐2‐hydroxy)benzylidene‐2‐(4‐oxoquinazolin‐3(4H)‐yl)acetohydrazides as antitumor agents

Dung

Park

Anh

et al. 2021

Archiv der Pharmazie

View full text Add to dashboard Cite

In our continuing search for novel small-molecule anticancer agents, we designed and synthesized a series of novel (E)-N'-(3-allyl-2-hydroxy)benzylidene-2-(4oxoquinazolin-3(4H)-yl)acetohydrazides ( 5), focusing on the modification of substitution in the quinazolin-4(3H)-one moiety. The biological evaluation showed that all 13 designed and synthesized compounds displayed significant cytotoxicity against three human cancer cell lines (SW620, colon cancer; PC-3, prostate cancer; NCI-H23, lung cancer). The most potent compound 5l displayed cytotoxicity up to 213fold more potent than 5-fluorouracil and 87-fold more potent than PAC-1, the first procaspase-activating compound. Structure-activity relationship analysis revealed that substitution of either electron-withdrawing or electron-releasing groups at positions 6 or 7 on the quinazolin-4(3H)-4-one moiety increased the cytotoxicity of the compounds, but substitution at position 6 seemed to be more favorable. In the caspase activation assay, compound 5l was found to activate the caspase activity by 291% in comparison to PAC-1, which was used as a control. Further docking simulation also revealed that this compound may be a potent allosteric inhibitor of procaspase-3 through chelation of the inhibitory zinc ion. Physicochemical and ADMET calculations for 5l provided useful information of its suitable absorption profile and some toxicological effects that need further optimization to be developed as a promising anticancer agent.

show abstract

Exploring different strategies for imbalanced ADME data problem: case study on Caco-2 permeability modeling

Cited by 11 publications

References 42 publications

Structure–activity relationship-based chemical classification of highly imbalanced Tox21 datasets

Structure–activity relationship-based chemical classification of highly imbalanced Tox21 datasets

Global Analysis of Models for Predicting Human Absorption: QSAR, In Vitro, and Preclinical Models

Design, synthesis, and evaluation of novel (E)‐N'‐(3‐allyl‐2‐hydroxy)benzylidene‐2‐(4‐oxoquinazolin‐3(4H)‐yl)acetohydrazides as antitumor agents

Contact Info

Product

Resources

About