How to approach machine learning-based prediction of drug/compound–target interactions

Abstract

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.

Exploring data‐driven chemical SMILES tokenization approaches to identify key protein–ligand binding moieties

“…ProtBENCH [30] contains protein family‐specific bioactivity data, spanning multiple protein superfamilies, including membrane receptors, ion channels, transporters, transcription factors, epigenetic regulators, and enzymes with five subgroups (i. e., transferases, proteases, hydrolases, oxidoreductases, and other enzymes). The family subsets vary in number of interactions (19 K—220 K), number of proteins (100—1 K), and number of compounds (10 K—120 K).…”

MocFormer: A Two-Stage Pre-training-Driven Transformer for Drug-Target Interactions Prediction

“…The pre-trained model from the DrugBank dataset underwent fine-tuning using the random forest regression method, and the learning rate was selected from the range [1e-5, 1e-4, 4e-4, 1e-3]. Furthermore, different batch sizes, namely [8, 16, 32], were experimented with. To ensure robustness, the five-fold cross-validation technique was utilized.…”

“…This approach allowed for the generation of negative samples, resulting in a balanced dataset for analysis. Dissimilar-compound-split dataset: This dataset is based on protein familyspecific datasets (Large-scale) [32], further constructed by applying a strategy that only considers compound similarities while distributing bioactivity data points into train-test splits, as presented in Table 2. Compounds in train and test splits are dissimilar (Tanimoto score < 0.5).…”

Redefining the Game: MVAE-DFDPnet’s Low-Dimensional Embeddings for Superior Drug-Protein Interaction Predictions

“…While the ligand-based approach relies on a sufficient number of known ligands for a given protein; the Molecular docking approach is limited to available 3D protein structures [8]. Conversely, machine learning-based methods have emerged as a highly promising avenue for predicting DPIs [9]- [12].…”