Identifying Biological Pathway Interrupting Toxins Using Multi-Tree Ensembles

Barta, Gergo

doi:10.3389/fenvs.2016.00052

Cited by 21 publications

(20 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, the RF classifier was widely used by the participating teams in the Tox21 Data Challenge [28, 48]. Two of the winning teams developed RF models that achieved the best performance in predicting compound activities against AR, aromatase, and p53 [ 58 ] as well as ER-LBD [ 59 ]. Using the same RF classifier and the same dataset made it convenient to compare our results with those from the participating teams and allowed us to better investigate the impact of resampling methods on improving imbalanced learning and, consequently, improving classification performance (see “ Comparison with Tox21 Data Challenge winners ” section below for more info).…”

Section: Resultsmentioning

confidence: 99%

“…Amaziz [ 68 ] employed associative neural networks to develop winning models for SR-ATAD5 and SR-MMP assays, and had the best overall BA score. Dmlab [ 58 ] used multi-tree ensemble methods, such as Random Forests and Extra Trees, to produce winning models for three assays (i.e., NR-AR, NR-aromatase and SR-p53). Microsomes [ 59 ] chose Random Forest for descriptor selection and model generation, and produced the best performing NR-ER-LBD model.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

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: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

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

“…[85] 0.824 AUC Best non-DNN model (multitree ensemble model) was placed third in the Tox21 challenge. [140] 0.838 AUC Runner up in Tox21 challenge was based off associative neural networks (ASNN). [141] 0.818 AUC Post-competition MT-DNN model.…”

Section: Non-dnn Models Commentsmentioning

confidence: 99%

Deep learning for computational chemistry

2017

View full text Add to dashboard Cite

The rise and fall of artificial neural networks is well documented in the scientific literature of both computer science and computational chemistry. Yet almost two decades later, we are now seeing a resurgence of interest in deep learning, a machine learning algorithm based on multilayer neural networks. Within the last few years, we have seen the transformative impact of deep learning in many domains, particularly in speech recognition and computer vision, to the extent that the majority of expert practitioners in those field are now regularly eschewing prior established models in favor of deep learning models. In this review, we provide an introductory overview into the theory of deep neural networks and their unique properties that distinguish them from traditional machine learning algorithms used in cheminformatics. By providing an overview of the variety of emerging applications of deep neural networks, we highlight its ubiquity and broad applicability to a wide range of challenges in the field, including QSAR, virtual screening, protein structure prediction, quantum chemistry, materials design and property prediction. In reviewing the performance of deep neural networks, we observed a consistent outperformance against nonneural networks state-of-the-art models across disparate research topics, and deep neural network based models often exceeded the "glass ceiling" expectations of their respective tasks. Coupled with the maturity of GPU-accelerated computing for training deep neural networks and the exponential growth of chemical data on which to train these networks on, we anticipate that deep learning algorithms will be a valuable tool for computational chemistry. 3 IntroductionDeep Learning is the key algorithm used in the development of AlphaGo, a Deep learning is a machine learning algorithm, not unlike those already in use in various applications in computational chemistry, from computer-aided drug design to materials property prediction. 5-8 Amongst some of its more high profile achievements include the Merck activity prediction challenge in 2012, where a deep neural network not only won the competition and outperformed Merck's internal baseline model, but did so without having a single chemist or biologist in their team. In a repeated success by a different research team, deep learning models achieved top positions in the Tox21 toxicity prediction challenge issued by NIH in 2014. 9 The unusually stellar performance of deep learning models in both predicting activity and toxicity in these recent examples, originate from the unique characteristics that distinguishes deep learning from traditional machine learning algorithms.For those unfamiliar with the intricacies of machine learning algorithms, we will highlight some of the key differences between traditional (shallow) machine learning and deep learning.The simplest example of a machine learning algorithm would be the ubiquitous least-squares linear regression. In linear regression, the underlying nature of the model is known (linear in th...

show abstract

“…The challenge leaderboard reports 10 models, ranked by their AUC (see also https ://tripo d.nih.gov/ tox21 /chall enge/leade rboar d.jsp). The winning model was an Extra trees classifier [29]. The models on ranks 2 to 6 were deep learning models [30].…”

Section: Comparison To the Tox21 Challenge Top10mentioning

confidence: 99%

COVER: conformational oversampling as data augmentation for molecules

2020

View full text Add to dashboard Cite

Training neural networks with small and imbalanced datasets often leads to overfitting and disregard of the minority class. For predictive toxicology, however, models with a good balance between sensitivity and specificity are needed. In this paper we introduce conformational oversampling as a means to balance and oversample datasets for prediction of toxicity. Conformational oversampling enhances a dataset by generation of multiple conformations of a molecule. These conformations can be used to balance, as well as oversample a dataset, thereby increasing the dataset size without the need of artificial samples. We show that conformational oversampling facilitates training of neural networks and provides state-of-the-art results on the Tox21 dataset.

show abstract

Identifying Biological Pathway Interrupting Toxins Using Multi-Tree Ensembles

Cited by 21 publications

References 17 publications

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

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

Deep learning for computational chemistry

COVER: conformational oversampling as data augmentation for molecules

Contact Info

Product

Resources

About