Comparison of Multilabel and Single-Label Classification Applied to the Prediction of the Isoform Specificity of Cytochrome P450 Substrates

Michielan, Lisa; Terfloth, Lothar; Gasteiger, Johann; Moro, Stefano

doi:10.1021/ci900299a

Cited by 50 publications

(42 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In reality a significant proportion of xenobiotics are metabolized by more than one isoform. This was addressed in a recent study employing multi-label classifiers such as SVMs, multilabel k-NNs and ANNs [278] based on a set of 580 substrates of seven P450 isoforms (approx. 15% of these substrates are known to bind to up to five enzyme variants).…”

Section: Modeling Studiesmentioning

confidence: 99%

The Challenges Involved in Modeling Toxicity Data In Silico: A Review

Gleeson¹,

Modi²,

Bender³

et al. 2012

CPD

View full text Add to dashboard Cite

The percentage of failures in late pharmaceutical development due to toxicity has increased dramatically over the last decade or so, resulting in increased demand for new methods to rapidly and reliably predict the toxicity of compounds. In this review we discuss the challenges involved in both the building of in silico models on toxicology endpoints and their practical use in decision making. In particular, we will reflect upon the predictive strength of a number of different in silico models for a range of different endpoints, different approaches used to generate the models or rules, and limitations of the methods and the data used in model generation. Given that there exists no unique definition of a 'good' model, we will furthermore highlight the need to balance model complexity/interpretability with predictability, particularly in light of OECD/REACH guidelines. Special emphasis is put on the data and methods used to generate the in silico toxicology models, and their strengths and weaknesses are discussed. Switching to the applied side, we next review a number of toxicity endpoints, discussing the methods available to predict them and their general level of predictability (which very much depends on the endpoint considered). We conclude that, while in silico toxicology is a valuable tool to drug discovery scientists, much still needs to be done to, firstly, understand more completely the biological mechanisms for toxicity and, secondly, to generate more rapid in vitro models to screen compounds. With this biological understanding, and additional data available, our ability to generate more predictive in silico models should significantly improve in the future.

show abstract

Section: Modeling Studiesmentioning

confidence: 99%

The Challenges Involved in Modeling Toxicity Data In Silico: A Review

Gleeson¹,

Modi²,

Bender³

et al. 2012

CPD

View full text Add to dashboard Cite

show abstract

“…In order to realize a better understanding of the function of chloroplast proteins, a proposed MLC algorithm was applied in prediction of protein subchloroplast locations in chloroplast organelle [27]. It was also shown that when the MLC approach is compared to a single label classification, it coherently reflects the actual metabolism information when applied over a collection of CYP450 substrates [30]. Multi-label Naïve Bayes classification models were constructed to improve target prediction for relevant target proteins over a wide set of chemical compounds [31].…”

Section: Introductionmentioning

confidence: 99%

DRABAL: novel method to mine large high-throughput screening assays using Bayesian active learning

et al. 2016

View full text Add to dashboard Cite

Background Mining high-throughput screening (HTS) assays is key for enhancing decisions in the area of drug repositioning and drug discovery. However, many challenges are encountered in the process of developing suitable and accurate methods for extracting useful information from these assays. Virtual screening and a wide variety of databases, methods and solutions proposed to-date, did not completely overcome these challenges. This study is based on a multi-label classification (MLC) technique for modeling correlations between several HTS assays, meaning that a single prediction represents a subset of assigned correlated labels instead of one label. Thus, the devised method provides an increased probability for more accurate predictions of compounds that were not tested in particular assays.ResultsHere we present DRABAL, a novel MLC solution that incorporates structure learning of a Bayesian network as a step to model dependency between the HTS assays. In this study, DRABAL was used to process more than 1.4 million interactions of over 400,000 compounds and analyze the existing relationships between five large HTS assays from the PubChem BioAssay Database. Compared to different MLC methods, DRABAL significantly improves the F1Score by about 22%, on average. We further illustrated usefulness and utility of DRABAL through screening FDA approved drugs and reported ones that have a high probability to interact with several targets, thus enabling drug-multi-target repositioning. Specifically DRABAL suggests the Thiabendazole drug as a common activator of the NCP1 and Rab-9A proteins, both of which are designed to identify treatment modalities for the Niemann–Pick type C disease.ConclusionWe developed a novel MLC solution based on a Bayesian active learning framework to overcome the challenge of lacking fully labeled training data and exploit actual dependencies between the HTS assays. The solution is motivated by the need to model dependencies between existing experimental confirmatory HTS assays and improve prediction performance. We have pursued extensive experiments over several HTS assays and have shown the advantages of DRABAL. The datasets and programs can be downloaded from https://figshare.com/articles/DRABAL/3309562.Graphical abstract. Electronic supplementary materialThe online version of this article (doi:10.1186/s13321-016-0177-8) contains supplementary material, which is available to authorized users.

show abstract

“…21 There are two main types of multilabel methods: problem transformation and algorithm adaption methods. Problem transformation methods involve transformation of the multilabel data into single label data to then carry out conventional single label classification.…”

Section: Introductionmentioning

confidence: 99%

Comparing Multilabel Classification Methods for Provisional Biopharmaceutics Class Prediction

2014

View full text Add to dashboard Cite

The biopharmaceutical classification system (BCS) is now well established and utilized for the development and biowaivers of immediate oral dosage forms. The prediction of BCS class can be carried out using multilabel classification. Unlike single label classification, multilabel classification methods predict more than one class label at the same time. This paper compares two multilabel methods, binary relevance and classifier chain, for provisional BCS class prediction. Large data sets of permeability and solubility of drug and drug-like compounds were obtained from the literature and were used to build models using decision trees. The separate permeability and solubility models were validated, and a BCS validation set of 127 compounds where both permeability and solubility were known was used to compare the two aforementioned multilabel classification methods for provisional BCS class prediction. Overall, the results indicate that the classifier chain method, which takes into account label interactions, performed better compared to the binary relevance method. This work offers a comparison of multilabel methods and shows the potential of the classifier chain multilabel method for improved biological property predictions for use in drug discovery and development.

show abstract

Comparison of Multilabel and Single-Label Classification Applied to the Prediction of the Isoform Specificity of Cytochrome P450 Substrates

Cited by 50 publications

References 46 publications

The Challenges Involved in Modeling Toxicity Data In Silico: A Review

The Challenges Involved in Modeling Toxicity Data In Silico: A Review

DRABAL: novel method to mine large high-throughput screening assays using Bayesian active learning

Comparing Multilabel Classification Methods for Provisional Biopharmaceutics Class Prediction

Contact Info

Product

Resources

About