Machine Learning for Sequence and Structure-Based Protein-Ligand Interaction Prediction

Developing new drugs is too expensive and time -consuming. Accurately predicting the interaction between drugs and targets will likely change how the drug is discovered. Machine learning-based protein-ligand interaction prediction has demonstrated significant potential. In this paper, computational methods, focusing on sequence and structure to study protein-ligand interactions, are examined. Therefore, this paper starts by presenting an overview of the data sets applied in this area, as well as the various approaches applied for representing proteins and ligands. Then, sequence-based and structure-based classification criteria are subsequently utilized to categorize and summarize both the classical machine learning models and deep learning models employed in protein-ligand interaction studies. Moreover, the evaluation methods and interpretability of these models are proposed. Furthermore, delving into the diverse applications of protein-ligand interaction models in drug research is presented. Lastly, the current challenges and future directions in this field are addressed.

PASSerRank: Prediction of allosteric sites with learning to rank

Allostery plays a crucial role in regulating protein activity, making it a highly sought-after target in drug development. One of the major challenges in allosteric drug research is the identification of allosteric sites. In recent years, many computational models have been developed for accurate allosteric site prediction. Most of these models focus on designing a general rule that can be applied to pockets of proteins from various families. In this study, we present a new approach using the concept of Learning to Rank (LTR). The LTR model ranks pockets based on their relevance to allosteric sites, that is, how well a pocket meets the characteristics of known allosteric sites. After the training and validation on two datasets, the Allosteric Database (ASD) and CASBench, the LTR model was able to rank an allosteric pocket in the top three positions for 83.6% and 80.5% of test proteins, respectively. The model outperforms other common machine learning models with higher F1 scores (0.662 in ASD and 0.608 in CASBench) and Matthews correlation coefficients (0.645 in ASD and 0.589 in CASBench). The trained model is available on the PASSer platform (https://passer.smu.edu) to aid in drug discovery research.

Machine learning and protein allostery

The fundamental biological importance and complexity of allosterically regulated proteins stem from their central role in signal transduction and cellular processes. Recently, machine-learning approaches have been developed and actively deployed to facilitate theoretical and experimental studies of protein dynamics and allosteric mechanisms. In this review, we survey recent developments in applications of machine-learning methods for studies of allosteric mechanisms, prediction of allosteric effects and allostery-related physicochemical properties, and allosteric protein engineering. We also review the applications of machine-learning strategies for characterization of allosteric mechanisms and drug design targeting SARS-CoV-2. Continuous development and task-specific adaptation of machine-learning methods for protein allosteric mechanisms will have an increasingly important role in bridging a wide spectrum of data-intensive experimental and theoretical technologies.

Prediction of Protein Allosteric Signalling Pathways and Functional Residues Through Paths of Optimised Propensity

Allostery commonly refers to the mechanism that regulates protein activity through the binding of a molecule at a different, usually distal, site from the orthosteric site. The omnipresence of allosteric regulation in nature and its potential for drug design and screening render the study of allostery invaluable. Nevertheless, challenges remain as few computational methods are available to effectively predict allosteric sites, identify signalling pathways involved in allostery, or to aid with the design of suitable molecules targeting such sites. Recently, bond-to-bond propensity analysis has been shown successful at identifying allosteric sites for a large and diverse group of proteins from knowledge of the orthosteric sites and its ligands alone by using network analysis applied to energy-weighted atomistic protein graphs. To address the identification of signalling pathways, we propose here a method to compute and score paths of optimised propensity that link the orthosteric site with the identified allosteric sites, and identifies crucial residues that contribute to those paths. We showcase the approach with three well-studied allosteric proteins: h-Ras, caspase-1, and 3-phosphoinositide-dependent kinase-1 (PDK1). Key residues in both orthosteric and allosteric sites were identified and showed agreement with experimental results, and pivotal signalling residues along the pathway were also revealed, thus providing alternative targets for drug design. By using the computed path scores, we were also able to differentiate the activity of different allosteric modulators.

PASSer2.0: Accurate Prediction of Protein Allosteric Sites Through Automated Machine Learning

Allostery is a fundamental process in regulating protein activities. The discovery, design, and development of allosteric drugs demand better identification of allosteric sites. Several computational methods have been developed previously to predict allosteric sites using static pocket features and protein dynamics. Here, we define a baseline model for allosteric site prediction and present a computational model using automated machine learning. Our model, PASSer2.0, advanced the previous results and performed well across multiple indicators with 82.7% of allosteric pockets appearing among the top three positions. The trained machine learning model has been integrated with the Protein Allosteric Sites Server (PASSer) to facilitate allosteric drug discovery.

Artificial intelligence in the prediction of protein-ligand interactions: recent advances and future directions

New drug production, from target identification to marketing approval, takes over 12 years and can cost around $2.6 billion. Furthermore, the COVID-19 pandemic has unveiled the urgent need for more powerful computational methods for drug discovery. Here, we review the computational approaches to predicting protein-ligand interactions in the context of drug discovery, focusing on methods using artificial intelligence (AI). We begin with a brief introduction to proteins (targets), ligands (e.g. drugs) and their interactions for nonexperts. Next, we review databases that are commonly used in the domain of protein-ligand interactions. Finally, we survey and analyze the machine learning (ML) approaches implemented to predict protein-ligand binding sites, ligand-binding affinity and binding pose (conformation) including both classical ML algorithms and recent deep learning methods. After exploring the correlation between these three aspects of protein-ligand interaction, it has been proposed that they should be studied in unison. We anticipate that our review will aid exploration and development of more accurate ML-based prediction strategies for studying protein-ligand interactions.

Prediction of allosteric sites and signaling: Insights from benchmarking datasets

Allostery is a pervasive mechanism that regulates protein activity through ligand binding at a site different from the orthosteric site. The universality of allosteric regulation complemented by the benefits of highly specific and potentially non-toxic allosteric drugs makes uncovering allosteric sites invaluable. However, there are few computational methods to effectively predict them. Bond-to-bond propensity analysis has successfully predicted allosteric sites in 19 of 20 cases using an energy-weighted atomistic graph. We here extended the analysis onto 432 structures of 146 proteins from two benchmarking datasets for allosteric proteins: ASBench and CASBench. We further introduced two statistical measures to account for the cumulative effect of high-propensity residues and the crucial residues in a given site. The allosteric site is recovered for 127 of 146 proteins (407 of 432 structures) knowing only the orthosteric sites or ligands. The quantitative analysis using a range of statistical measures enables better characterization of potential allosteric sites and mechanisms involved.

PASSer: Prediction of Allosteric Sites Server

Allostery is considered important in regulating protein's activity. Drug development depends on the understanding of allosteric mechanisms, especially the identification of allosteric sites, which is a prerequisite in drug discovery and design. Many computational methods have been developed for allosteric site prediction using pocket features and protein dynamics. Here, we present an ensemble learning method, consisting of eXtreme gradient boosting (XGBoost) and graph convolutional neural network (GCNN), to predict allosteric sites. Our model can learn physical properties and topology without any prior information, and shows good performance under multiple indicators. Prediction results showed that 84.9% of allosteric pockets in the test set appeared in the top 3 positions. The PASSer: Protein Allosteric Sites Server (https://passer.smu.edu), along with a command line interface (CLI, https://github.com/smutaogroup/passerCLI) provide insights for further analysis in drug discovery.

Drug design targeting active posttranslational modification protein isoforms

Modern drug design aims to discover novel lead compounds with attractable chemical profiles to enable further exploration of the intersection of chemical space and biological space. Identification of small molecules with good ligand efficiency, high activity, and selectivity is crucial toward developing effective and safe drugs. However, the intersection is one of the most challenging tasks in the pharmaceutical industry, as chemical space is almost infinity and continuous, whereas the biological space is very limited and discrete. This bottleneck potentially limits the discovery of molecules with desirable properties for lead optimization. Herein, we present a new direction leveraging posttranslational modification (PTM) protein isoforms target space to inspire drug design termed as "Post-translational Modification Inspired Drug Design (PTMI-DD)." PTMI-DD aims to extend the intersections of chemical space and biological space. We further rationalized and highlighted the importance of PTM protein isoforms and their roles in various diseases and biological functions. We then laid out a few directions to elaborate the PTMI-DD in drug design including discovering covalent binding inhibitors mimicking PTMs, targeting PTM protein isoforms with distinctive binding sites from that of wild-type counterpart, targeting protein-protein interactions involving PTMs, and hijacking protein degeneration by ubiquitination for PTM protein isoforms. These directions will lead to a significant expansion of the biological space and/or increase the tractability of compounds, primarily due to precisely targeting PTM protein isoforms or complexes which are highly relevant to biological functions. Importantly, this new avenue will further enrich the personalized treatment opportunity through precision medicine targeting PTM isoforms.

Structural dynamics and allostery of Rab proteins: strategies for drug discovery and design

Rab proteins represent the largest family of the Rab superfamily guanosine triphosphatase (GTPase). Aberrant human Rab proteins are associated with multiple diseases, including cancers and neurological disorders. Rab subfamily members display subtle conformational variations that render specificity in their physiological functions and can be targeted for subfamily-specific drug design. However, drug discovery efforts have not focused much on targeting Rab allosteric non-nucleotide binding sites which are subjected to less evolutionary pressures to be conserved, hence are likely to offer subfamily specificity and may be less prone to undesirable off-target interactions and side effects. To discover druggable allosteric binding sites, Rab structural dynamics need to be first incorporated using multiple experimentally and computationally obtained structures. The high-dimensional structural data may necessitate feature extraction methods to identify manageable representative structures for subsequent analyses. We have detailed state-of-the-art computational methods to (i) identify binding sites using data on sequence, shape, energy, etc., (ii) determine the allosteric nature of these binding sites based on structural ensembles, residue networks and correlated motions and (iii) identify small molecule binders through structure- and ligand-based virtual screening. To benefit future studies for targeting Rab allosteric sites, we herein detail a refined workflow comprising multiple available computational methods, which have been successfully used alone or in combinations. This workflow is also applicable for drug discovery efforts targeting other medically important proteins. Depending on the structural dynamics of proteins of interest, researchers can select suitable strategies for allosteric drug discovery and design, from the resources of computational methods and tools enlisted in the workflow.

AlloFinder: a strategy for allosteric modulator discovery and allosterome analyses

Allostery tweaks innumerable biological processes and plays a fundamental role in human disease and drug discovery. Exploration of allostery has thus been regarded as a crucial requirement for research on biological mechanisms and the development of novel therapeutics. Here, based on our previously developed allosteric data and methods, we present an interactive platform called AlloFinder that identifies potential endogenous or exogenous allosteric modulators and their involvement in human allosterome. AlloFinder automatically amalgamates allosteric site identification, allosteric screening and allosteric scoring evaluation of modulator-protein complexes to identify allosteric modulators, followed by allosterome mapping analyses of predicted allosteric sites and modulators in human proteome. This web server exhibits prominent performance in the reemergence of allosteric metabolites and exogenous allosteric modulators in known allosteric proteins. Specifically, AlloFinder enables identification of allosteric metabolites for metabolic enzymes and screening of potential allosteric compounds for disease-related targets. Significantly, the feasibility of AlloFinder to discover allosteric modulators was tested in a real case of signal transduction and activation of transcription 3 (STAT3) and validated by mutagenesis and functional experiments. Collectively, AlloFinder is expected to contribute to exploration of the mechanisms of allosteric regulation between metabolites and metabolic enzymes, and to accelerate allosteric drug discovery. The AlloFinder web server is freely available to all users at http://mdl.shsmu.edu.cn/ALF/.

Review: Precision medicine and driver mutations: Computational methods, functional assays and conformational principles for interpreting cancer drivers

At the root of the so-called precision medicine or precision oncology, which is our focus here, is the hypothesis that cancer treatment would be considerably better if therapies were guided by a tumor’s genomic alterations. This hypothesis has sparked major initiatives focusing on whole-genome and/or exome sequencing, creation of large databases, and developing tools for their statistical analyses—all aspiring to identify actionable alterations, and thus molecular targets, in a patient. At the center of the massive amount of collected sequence data is their interpretations that largely rest on statistical analysis and phenotypic observations. Statistics is vital, because it guides identification of cancer-driving alterations. However, statistics of mutations do not identify a change in protein conformation; therefore, it may not define sufficiently accurate actionable mutations, neglecting those that are rare. Among the many thematic overviews of precision oncology, this review innovates by further comprehensively including precision pharmacology, and within this framework, articulating its protein structural landscape and consequences to cellular signaling pathways. It provides the underlying physicochemical basis, thereby also opening the door to a broader community.

Allosteric Methods and Their Applications: Facilitating the Discovery of Allosteric Drugs and the Investigation of Allosteric Mechanisms

Allostery, or allosteric regulation, is the phenomenon in which protein functional activity is altered by the binding of an effector at an allosteric site that is topographically distinct from the orthosteric, active site. As one of the most direct and efficient ways to regulate protein function, allostery has played a fundamental role in innumerable biological processes of all living organisms, including enzyme catalysis, signal transduction, cell metabolism, and gene transcription. It is thus considered as "the second secret of life". The abnormality of allosteric communication networks between allosteric and orthosteric sites is associated with the pathogenesis of human diseases. Allosteric modulators, by attaching to structurally diverse allosteric sites, offer the potential for differential selectivity and improved safety compared with orthosteric drugs that bind to conserved orthosteric sites. Harnessing allostery has thus been regarded as a novel strategy for drug discovery. Despite much progress having been made in the repertoire of allostery since the turn of the millennium, the identification of allosteric drugs for therapeutic targets and the elucidation of allosteric mechanisms still present substantial challenges. These challenges are derived from the difficulties in the identification of allosteric sites and mutations, the assessment of allosteric protein-modulator interactions, the screening of allosteric modulators, and the elucidation of allosteric mechanisms in biological systems. To address these issues, we have developed a panel of allosteric services for specific allosteric applications over the past decade, including (i) the creation of the Allosteric Database, with the aim of providing comprehensive allosteric information such as allosteric proteins, modulators, sites, pathways, etc., (ii) the construction of the ASBench benchmark of high-quality allosteric sites for the development of computational methods for predicting allosteric sites, (iii) the development of Allosite and AllositePro for the prediction of the location of allosteric sites in proteins, (iv) the development of the Alloscore scoring function for the evaluation of allosteric protein-modulator interactions, (v) the development of Allosterome for evolutionary analysis of query allosteric sites/modulators within the human proteome, (vi) the development of AlloDriver for the prediction of allosteric mutagenesis, and (vii) the development of AlloFinder for the virtual screening of allosteric modulators and the investigation of allosteric mechanisms. Importantly, we have validated computationally predicted allosteric sites, mutations, and modulators in the real cases of sirtuin 6, casein kinase 2α, phosphodiesterase 10A, and signal transduction and activation of transcription 3. Furthermore, our developed allosteric methods have been widely exploited by other users around the world for allosteric research. Therefore, these allosteric services are expected to expedite the discovery of allosteric drugs and the investigation of allosteric mechanisms.

Advances in the Computational Identification of Allosteric Sites and Pathways in Proteins

With the increasing difficulty to develop new drugs and the emergence of resistance to traditional orthosteric-site inhibitors, the search for alternatives is finally approaching the focus on allosteric sites. Allosteric sites offer opportunities to regulate many pharmacologically targeted pathways by inhibition or activation. In addition, allosteric sites tend to be less conserved than the functional site, which may facilitate the design of specific effectors in the protein families for which specific orthosteric inhibitors have proved difficult to design. Furthermore, recent evidence suggests that all proteins might be susceptible of allosteric regulation, increasing the space of druggable targets. Computational identification of allosteric sites has therefore become an active field of research. The problem can be approached from two sides: (1) the identification of allosteric-communication pathways between the functional site and potential allosteric sites and (2) the functional-site-independent identification of allosteric sites. While the first approach tends to be more laborious and thus restricted to a single protein, the second tends to be more amenable to larger-scale analysis, thus providing tools for the two drug discovery scenarios: the analysis of known targets and the screening for new potential targets. Here, I show some basic concepts and methods useful to the identification of allosteric sites and pathways, in line with these two approaches. I describe them in some detail to build a clear framework, at the risk of losing the interest of experts. Examples of recent studies involving these methods are also illustrated, focusing on the techniques rather than on their findings on allosterism.

Drug Design for CNS Diseases: Polypharmacological Profiling of Compounds Using Cheminformatic, 3D-QSAR and Virtual Screening Methodologies

HIGHLIGHTS

Many CNS targets are being explored for multi-target drug design

New databases and cheminformatic methods enable prediction of primary pharmaceutical target and off-targets of compounds

QSAR, virtual screening and docking methods increase the potential of rational drug design

The diverse cerebral mechanisms implicated in Central Nervous System (CNS) diseases together with the heterogeneous and overlapping nature of phenotypes indicated that multitarget strategies may be appropriate for the improved treatment of complex brain diseases. Understanding how the neurotransmitter systems interact is also important in optimizing therapeutic strategies. Pharmacological intervention on one target will often influence another one, such as the well-established serotonin-dopamine interaction or the dopamine-glutamate interaction. It is now accepted that drug action can involve plural targets and that polypharmacological interaction with multiple targets, to address disease in more subtle and effective ways, is a key concept for development of novel drug candidates against complex CNS diseases. A multi-target therapeutic strategy for Alzheimer‘s disease resulted in the development of very effective Multi-Target Designed Ligands (MTDL) that act on both the cholinergic and monoaminergic systems, and also retard the progression of neurodegeneration by inhibiting amyloid aggregation. Many compounds already in databases have been investigated as ligands for multiple targets in drug-discovery programs. A probabilistic method, the Parzen-Rosenblatt Window approach, was used to build a “predictor” model using data collected from the ChEMBL database. The model can be used to predict both the primary pharmaceutical target and off-targets of a compound based on its structure. Several multi-target ligands were selected for further study, as compounds with possible additional beneficial pharmacological activities. Based on all these findings, it is concluded that multipotent ligands targeting AChE/MAO-A/MAO-B and also D₁-R/D₂-R/5-HT_2A-R/H₃-R are promising novel drug candidates with improved efficacy and beneficial neuroleptic and procognitive activities in treatment of Alzheimer's and related neurodegenerative diseases. Structural information for drug targets permits docking and virtual screening and exploration of the molecular determinants of binding, hence facilitating the design of multi-targeted drugs. The crystal structures and models of enzymes of the monoaminergic and cholinergic systems have been used to investigate the structural origins of target selectivity and to identify molecular determinants, in order to design MTDLs.