Iterative refinement of a binding pocket model: active computational steering of lead optimization

The present state and challenges of active learning in drug discovery

Active learning (AL) is an iterative feedback process that efficiently identifies valuable data within vast chemical space, even with limited labeled data. This characteristic renders it a valuable approach to tackle the ongoing challenges faced in drug discovery, such as the ever-expanding explore space and the limitations of labeled data. Consequently, AL is increasingly gaining prominence in the field of drug development. In this paper, we comprehensively review the application of AL at all stages of drug discovery, including compounds-target interaction prediction, virtual screening, molecular generation and optimization, as well as molecular properties prediction. Additionally, we discuss the challenges and prospects associated with the current applications of AL in drug discovery.

From UK-2A to florylpicoxamid: Active learning to identify a mimic of a macrocyclic natural product

Scaffold replacement as part of an optimization process that requires maintenance of potency, desirable biodistribution, metabolic stability, and considerations of synthesis at very large scale is a complex challenge. Here, we consider a set of over 1000 time-stamped compounds, beginning with a macrocyclic natural-product lead and ending with a broad-spectrum crop anti-fungal. We demonstrate the application of the QuanSA 3D-QSAR method employing an active learning procedure that combines two types of molecular selection. The first identifies compounds predicted to be most active of those most likely to be well-covered by the model. The second identifies compounds predicted to be most informative based on exhibiting low predicted activity but showing high 3D similarity to a highly active nearest-neighbor training molecule. Beginning with just 100 compounds, using a deterministic and automatic procedure, five rounds of 20-compound selection and model refinement identifies the binding metabolic form of florylpicoxamid. We show how iterative refinement broadens the domain of applicability of the successive models while also enhancing predictive accuracy. We also demonstrate how a simple method requiring very sparse data can be used to generate relevant ideas for synthetic candidates.

Benchmarking Active Learning Protocols for Ligand-Binding Affinity Prediction

Active learning (AL) has become a powerful tool in computational drug discovery, enabling the identification of top binders from vast molecular libraries. To design a robust AL protocol, it is important to understand the influence of AL parameters, as well as the features of the data sets on the outcomes. We use four affinity data sets for different targets (TYK2, USP7, D2R, Mpro) to systematically evaluate the performance of machine learning models [Gaussian process (GP) model and Chemprop model], sample selection protocols, and the batch size based on metrics describing the overall predictive power of the model (R2, Spearman rank, root-mean-square error) as well as the accurate identification of top 2%/5% binders (Recall, F1 score). Both models have a comparable Recall of top binders on large data sets, but the GP model surpasses the Chemprop model when training data are sparse. A larger initial batch size, especially on diverse data sets, increased the Recall of both models as well as overall correlation metrics. However, for subsequent cycles, smaller batch sizes of 20 or 30 compounds proved to be desirable. Furthermore, adding artificial Gaussian noise to the data up to a certain threshold still allowed the model to identify clusters with top-scoring compounds. However, excessive noise (<1σ) did impact the model's predictive and exploitative capabilities.

Novel Efficient Multistage Lead Optimization Pipeline Experimentally Validated for DYRK1B Selective Inhibitors

In addition to general challenges in drug discovery such as the identification of lead compounds in time- and cost-effective ways, specific challenges also exist. Particularly, it is necessary to develop pharmacological inhibitors that effectively discriminate between closely related molecular targets. DYRK1B kinase is considered a valuable target for cancer-specific mono- or combination chemotherapy; however, the inhibition of its closely related DYRK1A kinase is not beneficial. Existing inhibitors target both kinases with essentially the same efficiency, and the unavailability of the DYRK1B crystal structure makes the discovery of DYRK1B-specific inhibitors even more challenging. Here, we propose a novel multi-stage compound discovery pipeline aimed at in silico identification of both potent and selective small molecules from a large set of initial candidates. The method uses structure-based docking and ligand-based quantitative structure–activity relationship modeling. This approach allowed us to identify lead and runner-up small-molecule compounds targeting DYRK1B with high efficiency and specificity.

Data-driven discovery of cardiolipin-selective small molecules by computational active learning

Subtle variations in the lipid composition of mitochondrial membranes can have a profound impact on mitochondrial function. The inner mitochondrial membrane contains the phospholipid cardiolipin, which has been demonstrated to act as a biomarker for a number of diverse pathologies. Small molecule dyes capable of selectively partitioning into cardiolipin membranes enable visualization and quantification of the cardiolipin content. Here we present a data-driven approach that combines a deep learning-enabled active learning workflow with coarse-grained molecular dynamics simulations and alchemical free energy calculations to discover small organic compounds able to selectively permeate cardiolipin-containing membranes. By employing transferable coarse-grained models we efficiently navigate the all-atom design space corresponding to small organic molecules with molecular weight less than ≈500 Da. After direct simulation of only 0.42% of our coarse-grained search space we identify molecules with considerably increased levels of cardiolipin selectivity compared to a widely used cardiolipin probe 10-N-nonyl acridine orange. Our accumulated simulation data enables us to derive interpretable design rules linking coarse-grained structure to cardiolipin selectivity. The findings are corroborated by fluorescence anisotropy measurements of two compounds conforming to our defined design rules. Our findings highlight the potential of coarse-grained representations and multiscale modelling for materials discovery and design.

Synergy and Complementarity between Focused Machine Learning and Physics-Based Simulation in Affinity Prediction

We present results on the extent to which physics-based simulation (exemplified by FEP⁺) and focused machine learning (exemplified by QuanSA) are complementary for ligand affinity prediction. For both methods, predictions of activity for LFA-1 inhibitors from a medicinal chemistry lead optimization project were accurate within the applicable domain of each approach. A hybrid model that combined predictions by both approaches by simple averaging performed better than either method, with respect to both ranking and absolute pK_i values. Two publicly available FEP⁺ benchmarks, covering 16 diverse biological targets, were used to test the generality of the synergy. By identifying training data specifically focused on relevant ligands, accurate QuanSA models were derived using ligand activity data known at the time of the original series publications. Results across the 16 benchmark targets demonstrated significant improvements both for ranking and for absolute pK_i values using hybrid predictions that combined the FEP⁺ and QuanSA predicted affinity values. The results argue for a combined approach for affinity prediction that makes use of physics-driven methods as well as those driven by machine learning, each applied carefully on appropriate compounds, with hybrid prediction strategies being employed where possible.

Comprehensive strategies of machine-learning-based quantitative structure-activity relationship models

Early quantitative structure-activity relationship (QSAR) technologies have unsatisfactory versatility and accuracy in fields such as drug discovery because they are based on traditional machine learning and interpretive expert features. The development of Big Data and deep learning technologies significantly improve the processing of unstructured data and unleash the great potential of QSAR. Here we discuss the integration of wet experiments (which provide experimental data and reliable verification), molecular dynamics simulation (which provides mechanistic interpretation at the atomic/molecular levels), and machine learning (including deep learning) techniques to improve QSAR models. We first review the history of traditional QSAR and point out its problems. We then propose a better QSAR model characterized by a new iterative framework to integrate machine learning with disparate data input. Finally, we discuss the application of QSAR and machine learning to many practical research fields, including drug development and clinical trials.

Preliminary modelling as the first stage of targeted organic synthesis

The review aims to present a classification and applicability analysis of methods for preliminary molecular modelling for targeted organic, catalytic and biocatalytic synthesis. The following three main approaches are considered as a primary classification of the methods: modelling of the target – ligand coordination without structural information on both the target and the resulting complex; calculations based on experimentally obtained structural information about the target; and dynamic simulation of the target – ligand complex and the reaction mechanism with calculation of the free energy of the reaction. The review is meant for synthetic chemists to be used as a guide for building an algorithm for preliminary modelling and synthesis of structures with specified properties.

The bibliography includes 353 references.

Spectrum of deep learning algorithms in drug discovery

Deep learning (DL) algorithms are a subset of machine learning algorithms with the aim of modeling complex mapping between a set of elements and their classes. In parallel to the advance in revealing the molecular bases of diseases, a notable innovation has been undertaken to apply DL in data/libraries management, reaction optimizations, differentiating uncertainties, molecule constructions, creating metrics from qualitative results, and prediction of structures or interactions. From source identification to lead discovery and medicinal chemistry of the drug candidate, drug delivery, and modification, the challenges can be subjected to artificial intelligence algorithms to aid in the generation and interpretation of data. Discovery and design approach, both demand automation, large data management and data fusion by the advance in high-throughput mode. The application of DL can accelerate the exploration of drug mechanisms, finding novel indications for existing drugs (drug repositioning), drug development, and preclinical and clinical studies. The impact of DL in the workflow of drug discovery, design, and their complementary tools are highlighted in this review. Additionally, the type of DL algorithms used for this purpose, and their pros and cons along with the dominant directions of future research are presented.

Practical considerations for active machine learning in drug discovery

Active machine learning enables the automated selection of the most valuable next experiments to improve predictive modelling and hasten active retrieval in drug discovery. Although a long established theoretical concept and introduced to drug discovery approximately 15 years ago, the deployment of active learning technology in the discovery pipelines across academia and industry remains slow. With the recent re-discovered enthusiasm for artificial intelligence as well as improved flexibility of laboratory automation, active learning is expected to surge and become a key technology for molecular optimizations. This review recapitulates key findings from previous active learning studies to highlight the challenges and opportunities of applying adaptive machine learning to drug discovery. Specifically, considerations regarding implementation, infrastructural integration, and expected benefits are discussed. By focusing on these practical aspects of active learning, this review aims at providing insights for scientists planning to implement active learning workflows in their discovery pipelines.

Rational Drug Design of Antineoplastic Agents Using 3D-QSAR, Cheminformatic, and Virtual Screening Approaches

Background:

Computer-Aided Drug Design has strongly accelerated the development of novel antineoplastic agents by helping in the hit identification, optimization, and evaluation.

Results:

Computational approaches such as cheminformatic search, virtual screening, pharmacophore modeling, molecular docking and dynamics have been developed and applied to explain the activity of bioactive molecules, design novel agents, increase the success rate of drug research, and decrease the total costs of drug discovery. Similarity, searches and virtual screening are used to identify molecules with an increased probability to interact with drug targets of interest, while the other computational approaches are applied for the design and evaluation of molecules with enhanced activity and improved safety profile.

Conclusion:

In this review are described the main in silico techniques used in rational drug design of antineoplastic agents and presented optimal combinations of computational methods for design of more efficient antineoplastic drugs.

Iterative Screening Methods for Identification of Chemical Compounds with Specific Values of Various Properties

Identification of chemical compounds having desirable properties is a central goal of screening campaigns. Iterative screening is a means of surveying a set of compounds, during which their property values are determined and used as feedback for regression models. Quantitative models that assess the relationships between chemical structures and property/activity are repeatedly updated through this type of cycle, and the efficient sampling of compounds for the subsequent test is a key factor in the early identification of target compounds. Nevertheless, methodological approaches to comparisons and to establishing the degree of extrapolation of sampled compounds, including the effects of applicability domains, are still required. In the present study, we conducted a series of virtual experiments to assess the characteristics of different iterative screening methods. Genetic algorithm-based partial least-squares regression, support vector regression, Bayesian optimization with Gaussian Process (GP), and batch-based Bayesian optimization with GP (GP_batch) were all compared, based on the analysis of one million compounds extracted from the ZINC database. Our results show that, irrespective of the diversity of the initial set of compounds, it was possible to identify a compound having the desired property value using the appropriate screening method. However, overall, the GP_batch method was found to be preferable when evaluating properties either which are difficult to predict or for which a key factor is present in the set of molecular descriptors.

Quantitative surface field analysis: learning causal models to predict ligand binding affinity and pose

We introduce the QuanSA method for inducing physically meaningful field-based models of ligand binding pockets based on structure-activity data alone. The method is closely related to the QMOD approach, substituting a learned scoring field for a pocket constructed of molecular fragments. The problem of mutual ligand alignment is addressed in a general way, and optimal model parameters and ligand poses are identified through multiple-instance machine learning. We provide algorithmic details along with performance results on sixteen structure-activity data sets covering many pharmaceutically relevant targets. In particular, we show how models initially induced from small data sets can extrapolatively identify potent new ligands with novel underlying scaffolds with very high specificity. Further, we show that combining predictions from QuanSA models with those from physics-based simulation approaches is synergistic. QuanSA predictions yield binding affinities, explicit estimates of ligand strain, associated ligand pose families, and estimates of structural novelty and confidence. The method is applicable for fine-grained lead optimization as well as potent new lead identification.

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.

Multi-objective active machine learning rapidly improves structure-activity models and reveals new protein-protein interaction inhibitors

Active machine learning puts artificial intelligence in charge of a sequential, feedback-driven discovery process. We present the application of a multi-objective active learning scheme for identifying small molecules that inhibit the protein-protein interaction between the anti-cancer target CXC chemokine receptor 4 (CXCR4) and its endogenous ligand CXCL-12 (SDF-1). Experimental design by active learning was used to retrieve informative active compounds that continuously improved the adaptive structure-activity model. The balanced character of the compound selection function rapidly delivered new molecular structures with the desired inhibitory activity and at the same time allowed us to focus on informative compounds for model adjustment. The results of our study validate active learning for prospective ligand finding by adaptive, focused screening of large compound repositories and virtual compound libraries.

Extrapolative prediction using physically-based QSAR

Surflex-QMOD integrates chemical structure and activity data to produce physically-realistic models for binding affinity prediction . Here, we apply QMOD to a 3D-QSAR benchmark dataset and show broad applicability to a diverse set of targets. Testing new ligands within the QMOD model employs automated flexible molecular alignment, with the model itself defining the optimal pose for each ligand. QMOD performance was compared to that of four approaches that depended on manual alignments (CoMFA, two variations of CoMSIA, and CMF). QMOD showed comparable performance to the other methods on a challenging, but structurally limited, test set. The QMOD models were also applied to test a large and structurally diverse dataset of ligands from ChEMBL, nearly all of which were synthesized years after those used for model construction. Extrapolation across diverse chemical structures was possible because the method addresses the ligand pose problem and provides structural and geometric means to quantitatively identify ligands within a model’s applicability domain. Predictions for such ligands for the four tested targets were highly statistically significant based on rank correlation. Those molecules predicted to be highly active (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {pK}_i \ge 7.5$$\end{document}pKi≥7.5) had a mean experimental \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hbox {pK}_i$$\end{document}pKi of 7.5, with potent and structurally novel ligands being identified by QMOD for each target.

Is ring breaking feasible in relative binding free energy calculations?

Our interest is relative binding free energy (RBFE) calculations based on molecular simulations. These are promising tools for lead optimization in drug discovery, computing changes in binding free energy due to modifications of a lead compound. However, in the "alchemical" framework for RBFE calculations, some types of mutations have the potential to introduce error into the computed binding free energies. Here we explore the magnitude of this error in several different model binding calculations. We find that some of the calculations involving ring breaking have significant errors, and these errors are especially large in bridged ring systems. Since the error is a function of ligand strain, which is unpredictable in advance, we believe that ring breaking should be avoided when possible.

QSAR modeling: where have you been? Where are you going to?

Quantitative structure-activity relationship modeling is one of the major computational tools employed in medicinal chemistry. However, throughout its entire history it has drawn both praise and criticism concerning its reliability, limitations, successes, and failures. In this paper, we discuss (i) the development and evolution of QSAR; (ii) the current trends, unsolved problems, and pressing challenges; and (iii) several novel and emerging applications of QSAR modeling. Throughout this discussion, we provide guidelines for QSAR development, validation, and application, which are summarized in best practices for building rigorously validated and externally predictive QSAR models. We hope that this Perspective will help communications between computational and experimental chemists toward collaborative development and use of QSAR models. We also believe that the guidelines presented here will help journal editors and reviewers apply more stringent scientific standards to manuscripts reporting new QSAR studies, as well as encourage the use of high quality, validated QSARs for regulatory decision making.

A structure-guided approach for protein pocket modeling and affinity prediction

Binding affinity prediction is frequently addressed using computational models constructed solely with molecular structure and activity data. We present a hybrid structure-guided strategy that combines molecular similarity, docking, and multiple-instance learning such that information from protein structures can be used to inform models of structure-activity relationships. The Surflex-QMOD approach has been shown to produce accurate predictions of binding affinity by constructing an interpretable physical model of a binding site with no experimental binding site structural information. We introduce a method to integrate protein structure information into the model induction process in order to construct more robust physical models. The structure-guided models accurately predict binding affinities over a broad range of compounds while producing more accurate representations of the protein pockets and ligand binding modes. Structure-guidance for the QMOD method yielded significant performance improvements, both for affinity and pose prediction, especially in cases where predictions were made on ligands very different from those used for model induction.