Dataset Information

Drug Target Identification with Machine Learning: How to Choose Negative Examples.

ABSTRACT: Identification of the protein targets of hit molecules is essential in the drug discovery process. Target prediction with machine learning algorithms can help accelerate this search, limiting the number of required experiments. However, Drug-Target Interactions databases used for training present high statistical bias, leading to a high number of false positives, thus increasing time and cost of experimental validation campaigns. To minimize the number of false positives among predicted targets, we propose a new scheme for choosing negative examples, so that each protein and each drug appears an equal number of times in positive and negative examples. We artificially reproduce the process of target identification for three specific drugs, and more globally for 200 approved drugs. For the detailed three drug examples, and for the larger set of 200 drugs, training with the proposed scheme for the choice of negative examples improved target prediction results: the average number of false positives among the top ranked predicted targets decreased, and overall, the rank of the true targets was improved.Our method corrects databases' statistical bias and reduces the number of false positive predictions, and therefore the number of useless experiments potentially undertaken.

SUBMITTER: Najm M

PROVIDER: S-EPMC8151112 | biostudies-literature | 2021 May

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Drug Target Identification with Machine Learning: How to Choose Negative Examples.

Najm Matthieu M Azencott Chloé-Agathe CA Playe Benoit B Stoven Véronique V

International journal of molecular sciences 20210512 10

Identification of the protein targets of hit molecules is essential in the drug discovery process. Target prediction with machine learning algorithms can help accelerate this search, limiting the number of required experiments. However, Drug-Target Interactions databases used for training present high statistical bias, leading to a high number of false positives, thus increasing time and cost of experimental validation campaigns. To minimize the number of false positives among predicted targets, ...[more]

PMID: 34066072

Similar Datasets

Project description:BackgroundThe application of machine learning to classification problems that depend only on positive examples is gaining attention in the computational biology community. We and others have described the use of two-class machine learning to identify novel miRNAs. These methods require the generation of an artificial negative class. However, designation of the negative class can be problematic and if it is not properly done can affect the performance of the classifier dramatically and/or yield a biased estimate of performance. We present a study using one-class machine learning for microRNA (miRNA) discovery and compare one-class to two-class approaches using naïve Bayes and Support Vector Machines. These results are compared to published two-class miRNA prediction approaches. We also examine the ability of the one-class and two-class techniques to identify miRNAs in newly sequenced species.ResultsOf all methods tested, we found that 2-class naive Bayes and Support Vector Machines gave the best accuracy using our selected features and optimally chosen negative examples. One class methods showed average accuracies of 70-80% versus 90% for the two 2-class methods on the same feature sets. However, some one-class methods outperform some recently published two-class approaches with different selected features. Using the EBV genome as and external validation of the method we found one-class machine learning to work as well as or better than a two-class approach in identifying true miRNAs as well as predicting new miRNAs.ConclusionOne and two class methods can both give useful classification accuracies when the negative class is well characterized. The advantage of one class methods is that it eliminates guessing at the optimal features for the negative class when they are not well defined. In these cases one-class methods can be superior to two-class methods when the features which are chosen as representative of that positive class are well defined.

Project description:BackgroundmiRNAs are ~21 nucleotide long small noncoding RNA molecules, formed endogenously in most of the eukaryotes, which mainly control their target genes post transcriptionally by interacting and silencing them. While a lot of tools has been developed for animal miRNA target system, plant miRNA target identification system has witnessed limited development. Most of them have been centered around exact complementarity match. Very few of them considered other factors like multiple target sites and role of flanking regions.ResultIn the present work, a Support Vector Regression (SVR) approach has been implemented for plant miRNA target identification, utilizing position specific dinucleotide density variation information around the target sites, to yield highly reliable result. It has been named as p-TAREF (plant-Target Refiner). Performance comparison for p-TAREF was done with other prediction tools for plants with utmost rigor and where p-TAREF was found better performing in several aspects. Further, p-TAREF was run over the experimentally validated miRNA targets from species like Arabidopsis, Medicago, Rice and Tomato, and detected them accurately, suggesting gross usability of p-TAREF for plant species. Using p-TAREF, target identification was done for the complete Rice transcriptome, supported by expression and degradome based data. miR156 was found as an important component of the Rice regulatory system, where control of genes associated with growth and transcription looked predominant. The entire methodology has been implemented in a multi-threaded parallel architecture in Java, to enable fast processing for web-server version as well as standalone version. This also makes it to run even on a simple desktop computer in concurrent mode. It also provides a facility to gather experimental support for predictions made, through on the spot expression data analysis, in its web-server version.ConclusionA machine learning multivariate feature tool has been implemented in parallel and locally installable form, for plant miRNA target identification. The performance was assessed and compared through comprehensive testing and benchmarking, suggesting a reliable performance and gross usability for transcriptome wide plant miRNA target identification.

Project description:The identification of drug/compound-target interactions (DTIs) constitutes the basis of drug discovery, for which computational predictive approaches have been developed. As a relatively new data-driven paradigm, proteochemometric (PCM) modeling utilizes both protein and compound properties as a pair at the input level and processes them via statistical/machine learning. The representation of input samples (i.e., proteins and their ligands) in the form of quantitative feature vectors is crucial for the extraction of interaction-related properties during the artificial learning and subsequent prediction of DTIs. Lately, the representation learning approach, in which input samples are automatically featurized via training and applying a machine/deep learning model, has been utilized in biomedical sciences. In this study, we performed a comprehensive investigation of different computational approaches/techniques for protein featurization (including both conventional approaches and the novel learned embeddings), data preparation and exploration, machine learning-based modeling, and performance evaluation with the aim of achieving better data representations and more successful learning in DTI prediction. For this, we first constructed realistic and challenging benchmark datasets on small, medium, and large scales to be used as reliable gold standards for specific DTI modeling tasks. We developed and applied a network analysis-based splitting strategy to divide datasets into structurally different training and test folds. Using these datasets together with various featurization methods, we trained and tested DTI prediction models and evaluated their performance from different angles. Our main findings can be summarized under 3 items: (i) random splitting of datasets into train and test folds leads to near-complete data memorization and produce highly over-optimistic results, as a result, should be avoided, (ii) learned protein sequence embeddings work well in DTI prediction and offer high potential, despite interaction-related properties (e.g., structures) of proteins are unused during their self-supervised model training, and (iii) during the learning process, PCM models tend to rely heavily on compound features while partially ignoring protein features, primarily due to the inherent bias in DTI data, indicating the requirement for new and unbiased datasets. We hope this study will aid researchers in designing robust and high-performing data-driven DTI prediction systems that have real-world translational value in drug discovery.

Dataset Information

Drug Target Identification with Machine Learning: How to Choose Negative Examples.

Publications

Drug Target Identification with Machine Learning: How to Choose Negative Examples.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets