Abstract:In the drug discovery process, where experiments can be costly and time-consuming, computational models that predict drug-target interactions are valuable tools to accelerate the development of new therapeutic agents. Estimating the uncertainty inherent in these neural network predictions provides valuable information that facilitates optimal decision-making when risk assessment is crucial. However, such models can be poorly calibrated, which results in unreliable uncertainty estimates that do not reflect the true predictive uncertainty. In this study, we compare different metrics, including accuracy and calibration scores, used for model hyperparameter tuning to investigate which model selection strategy achieves well-calibrated models. Furthermore, we propose to use a computationally efficient Bayesian uncertainty estimation method named Bayesian Linear Probing (BLP), which generates Hamiltonian Monte Carlo (HMC) trajectories to obtain samples for the parameters of a Bayesian Logistic Regression fitted to the hidden layer of the baseline neural network. We report that BLP improves model calibration and achieves the performance of common uncertainty quantification methods by combining the benefits of uncertainty estimation and probability calibration methods. Finally, we show that combining post hoc calibration method with well-performing uncertainty quantification approaches can boost model accuracy and calibration.
Abstract:Deep learning-based approaches for generating novel drug molecules with specific properties have gained a lot of interest in the last years. Recent studies have demonstrated promising performance for string-based generation of novel molecules utilizing reinforcement learning. In this paper, we develop a unified framework for using reinforcement learning for de novo drug design, wherein we systematically study various on- and off-policy reinforcement learning algorithms and replay buffers to learn an RNN-based policy to generate novel molecules predicted to be active against the dopamine receptor DRD2. Our findings suggest that it is advantageous to use at least both top-scoring and low-scoring molecules for updating the policy when structural diversity is essential. Using all generated molecules at an iteration seems to enhance performance stability for on-policy algorithms. In addition, when replaying high, intermediate, and low-scoring molecules, off-policy algorithms display the potential of improving the structural diversity and number of active molecules generated, but possibly at the cost of a longer exploration phase. Our work provides an open-source framework enabling researchers to investigate various reinforcement learning methods for de novo drug design.
Abstract:To apply federated learning to drug discovery we developed a novel platform in the context of European Innovative Medicines Initiative (IMI) project MELLODDY (grant n{\deg}831472), which was comprised of 10 pharmaceutical companies, academic research labs, large industrial companies and startups. To the best of our knowledge, The MELLODDY platform was the first industry-scale platform to enable the creation of a global federated model for drug discovery without sharing the confidential data sets of the individual partners. The federated model was trained on the platform by aggregating the gradients of all contributing partners in a cryptographic, secure way following each training iteration. The platform was deployed on an Amazon Web Services (AWS) multi-account architecture running Kubernetes clusters in private subnets. Organisationally, the roles of the different partners were codified as different rights and permissions on the platform and administrated in a decentralized way. The MELLODDY platform generated new scientific discoveries which are described in a companion paper.
Abstract:Recent developments in artificial intelligence and automation could potentially enable a new drug design paradigm: autonomous drug design. Under this paradigm, generative models provide suggestions on thousands of molecules with specific properties. However, since only a limited number of molecules can be synthesized and tested, an obvious challenge is how to efficiently select these. We formulate this task as a contextual stochastic multi-armed bandit problem with multiple plays and volatile arms. Then, to solve it, we extend previous work on multi-armed bandits to reflect this setting, and compare our solution with random sampling, greedy selection and decaying-epsilon-greedy selection. To investigate how the different selection strategies affect the cumulative reward and the diversity of the selections, we simulate the drug design process. According to the simulation results, our approach has the potential for better exploring and exploiting the chemical space for autonomous drug design.
Abstract:Improving on the standard of care for diseases is predicated on better treatments, which in turn relies on finding and developing new drugs. However, drug discovery is a complex and costly process. Adoption of methods from machine learning has given rise to creation of drug discovery knowledge graphs which utilize the inherent interconnected nature of the domain. Graph-based data modelling, combined with knowledge graph embeddings provide a more intuitive representation of the domain and are suitable for inference tasks such as predicting missing links. One such example would be producing ranked lists of likely associated genes for a given disease, often referred to as target discovery. It is thus critical that these predictions are not only pertinent but also biologically meaningful. However, knowledge graphs can be biased either directly due to the underlying data sources that are integrated or due to modeling choices in the construction of the graph, one consequence of which is that certain entities can get topologically overrepresented. We show how knowledge graph embedding models can be affected by this structural imbalance, resulting in densely connected entities being highly ranked no matter the context. We provide support for this observation across different datasets, models and predictive tasks. Further, we show how the graph topology can be perturbed to artificially alter the rank of a gene via random, biologically meaningless information. This suggests that such models can be more influenced by the frequency of entities rather than biological information encoded in the relations, creating issues when entity frequency is not a true reflection of underlying data. Our results highlight the importance of data modeling choices and emphasizes the need for practitioners to be mindful of these issues when interpreting model outputs and during knowledge graph composition.
Abstract:Capsule Networks (CapsNets) is a machine learning architecture proposed to overcome some of the shortcomings of convolutional neural networks (CNNs). However, CapsNets have mainly outperformed CNNs in datasets where images are small and/or the objects to identify have minimal background noise. In this work, we present a new architecture, parallel CapsNets, which exploits the concept of branching the network to isolate certain capsules, allowing each branch to identify different entities. We applied our concept to the two current types of CapsNet architectures, studying the performance for networks with different layers of capsules. We tested our design in a public, highly unbalanced dataset of acute myeloid leukaemia images (15 classes). Our experiments showed that conventional CapsNets show similar performance than our baseline CNN (ResNeXt-50) but depict instability problems. In contrast, parallel CapsNets can outperform ResNeXt-50, is more stable, and shows better rotational invariance than both, conventional CapsNets and ResNeXt-50.
Abstract:Knowledge Graphs (KG) and associated Knowledge Graph Embedding (KGE) models have recently begun to be explored in the context of drug discovery and have the potential to assist in key challenges such as target identification. In the drug discovery domain, KGs can be employed as part of a process which can result in lab-based experiments being performed, or impact on other decisions, incurring significant time and financial costs and most importantly, ultimately influencing patient healthcare. For KGE models to have impact in this domain, a better understanding of not only of performance, but also the various factors which determine it, is required. In this study we investigate, over the course of many thousands of experiments, the predictive performance of five KGE models on two public drug discovery-oriented KGs. Our goal is not to focus on the best overall model or configuration, instead we take a deeper look at how performance can be affected by changes in the training setup, choice of hyperparameters, model parameter initialisation seed and different splits of the datasets. Our results highlight that these factors have significant impact on performance and can even affect the ranking of models. Indeed these factors should be reported along with model architectures to ensure complete reproducibility and fair comparisons of future work, and we argue this is critical for the acceptance of use, and impact of KGEs in a biomedical setting. To aid reproducibility of our own work, we release all experimentation code.
Abstract:Drug discovery and development is an extremely complex process, with high attrition contributing to the costs of delivering new medicines to patients. Recently, various machine learning approaches have been proposed and investigated to help improve the effectiveness and speed of multiple stages of the drug discovery pipeline. Among these techniques, it is especially those using Knowledge Graphs that are proving to have considerable promise across a range of tasks, including drug repurposing, drug toxicity prediction and target gene-disease prioritisation. In such a knowledge graph-based representation of drug discovery domains, crucial elements including genes, diseases and drugs are represented as entities or vertices, whilst relationships or edges between them indicate some level of interaction. For example, an edge between a disease and drug entity might represent a successful clinical trial, or an edge between two drug entities could indicate a potentially harmful interaction. In order to construct high-quality and ultimately informative knowledge graphs however, suitable data and information is of course required. In this review, we detail publicly available primary data sources containing information suitable for use in constructing various drug discovery focused knowledge graphs. We aim to help guide machine learning and knowledge graph practitioners who are interested in applying new techniques to the drug discovery field, but who may be unfamiliar with the relevant data sources. Overall we hope this review will help motivate more machine learning researchers to explore combining knowledge graphs and machine learning to help solve key and emerging questions in the drug discovery domain.
Abstract:A major challenge in computational chemistry is the generation of novel molecular structures with desirable pharmacological and physiochemical properties. In this work, we investigate the potential use of autoencoder, a deep learning methodology, for de novo molecular design. Various generative autoencoders were used to map molecule structures into a continuous latent space and vice versa and their performance as structure generator was assessed. Our results show that the latent space preserves chemical similarity principle and thus can be used for the generation of analogue structures. Furthermore, the latent space created by autoencoders were searched systematically to generate novel compounds with predicted activity against dopamine receptor type 2 and compounds similar to known active compounds not included in the training set were identified.
Abstract:This work introduces a method to tune a sequence-based generative model for molecular de novo design that through augmented episodic likelihood can learn to generate structures with certain specified desirable properties. We demonstrate how this model can execute a range of tasks such as generating analogues to a query structure and generating compounds predicted to be active against a biological target. As a proof of principle, the model is first trained to generate molecules that do not contain sulphur. As a second example, the model is trained to generate analogues to the drug Celecoxib, a technique that could be used for scaffold hopping or library expansion starting from a single molecule. Finally, when tuning the model towards generating compounds predicted to be active against the dopamine receptor type 2, the model generates structures of which more than 95% are predicted to be active, including experimentally confirmed actives that have not been included in either the generative model nor the activity prediction model.