Augmenting DMTA using predictive AI modelling at AstraZeneca.

Design-Make-Test-Analyse (DMTA) is the discovery cycle through which molecules are designed, synthesised, and assayed to produce data that in turn are analysed to inform the next iteration. The process is repeated until viable drug candidates are identified, often requiring many cycles before reaching a sweet spot. The advent of artificial intelligence (AI) and cloud computing presents an opportunity to innovate drug discovery to reduce the number of cycles needed to yield a candidate. Here, we present the Predictive Insight Platform (PIP), a cloud-native modelling platform developed at AstraZeneca. The impact of PIP in each step of DMTA, as well as its architecture, integration, and usage, are discussed and used to provide insights into the future of drug discovery.

Practical High-Quality Electrostatic Potential Surfaces for Drug Discovery Using a Graph-Convolutional Deep Neural Network.

Inspecting protein and ligand electrostatic potential (ESP) surfaces in order to optimize electrostatic complementarity is a key activity in drug design. These ESP surfaces need to reflect the true electrostatic nature of the molecules, which typically means time-consuming high-level quantum mechanics (QM) calculations are required. For interactive design much faster alternative methods are required. Here, we present a graph convolutional deep neural network (DNN) model, trained on ESP surfaces derived from high quality QM calculations, that generates ESP surfaces for ligands in a fraction of a second. Additionally, we describe a method for constructing fast QM-trained ESP surfaces for proteins. We show that the DNN model generates ESP surfaces that are in good agreement with QM and that the ESP values correlate well with experimental properties relevant to medicinal chemistry. We believe that these high-quality, interactive ESP surfaces form a powerful tool for driving drug discovery programs forward. The trained model and associated code are available from https://github.com/AstexUK/ESP_DNN.

Predicting "Hot" and "Warm" Spots for Fragment Binding.

Computational fragment mapping methods aim to predict hotspots on protein surfaces where small fragments will bind. Such methods are popular for druggability assessment as well as structure-based design. However, to date researchers developing or using such tools have had no clear way of assessing the performance of these methods. Here, we introduce the first diverse, high quality validation set for computational fragment mapping. The set contains 52 diverse examples of fragment binding "hot" and "warm" spots from the Protein Data Bank (PDB). Additionally, we describe PLImap, a novel protocol for fragment mapping based on the Protein-Ligand Informatics force field (PLIff). We evaluate PLImap against the new fragment mapping test set, and compare its performance to that of simple shape-based algorithms and fragment docking using GOLD. PLImap is made publicly available from https://bitbucket.org/AstexUK/pli .

Protein-Ligand Informatics Force Field (PLIff): Toward a Fully Knowledge Driven "Force Field" for Biomolecular Interactions.

The Protein Data Bank (PDB) contains a wealth of data on nonbonded biomolecular interactions. If this information could be distilled down to nonbonded interaction potentials, these would have some key advantages over standard force fields. However, there are some important outstanding issues to address in order to do this successfully. This paper introduces the protein-ligand informatics "force field", PLIff, which begins to address these key challenges ( https://bitbucket.org/AstexUK/pli ). As a result of their knowledge-based nature, the next-generation nonbonded potentials that make up PLIff automatically capture a wide range of interaction types, including special interactions that are often poorly described by standard force fields. We illustrate how PLIff may be used in structure-based design applications, including interaction fields, fragment mapping, and protein-ligand docking. PLIff performs at least as well as state-of-the art scoring functions in terms of pose predictions and ranking compounds in a virtual screening context.

Application of Rigidity Theory to the Thermostabilization of Lipase A from Bacillus subtilis.

Protein thermostability is a crucial factor for biotechnological enzyme applications. Protein engineering studies aimed at improving thermostability have successfully applied both directed evolution and rational design. However, for rational approaches, the major challenge remains the prediction of mutation sites and optimal amino acid substitutions. Recently, we showed that such mutation sites can be identified as structural weak spots by rigidity theory-based thermal unfolding simulations of proteins. Here, we describe and validate a unique, ensemble-based, yet highly efficient strategy to predict optimal amino acid substitutions at structural weak spots for improving a protein's thermostability. For this, we exploit the fact that in the majority of cases an increased structural rigidity of the folded state has been found as the cause for thermostability. When applied prospectively to lipase A from Bacillus subtilis, we achieved both a high success rate (25% over all experimentally tested mutations, which raises to 60% if small-to-large residue mutations and mutations in the active site are excluded) in predicting significantly thermostabilized lipase variants and a remarkably large increase in those variants' thermostability (up to 6.6°C) based on single amino acid mutations. When considering negative controls in addition and evaluating the performance of our approach as a binary classifier, the accuracy is 63% and increases to 83% if small-to-large residue mutations and mutations in the active site are excluded. The gain in precision (predictive value for increased thermostability) over random classification is 1.6-fold (2.4-fold). Furthermore, an increase in thermostability predicted by our approach significantly points to increased experimental thermostability (p < 0.05). These results suggest that our strategy is a valuable complement to existing methods for rational protein design aimed at improving thermostability.

Design and synthesis of novel Y-shaped barbituric acid derivatives as PPARγ activators.

Novel Y-shaped barbituric acid (BA) derivatives have been designed using rational methods including molecular docking. Fourteen novel compounds were synthesized using hydroxyl group protection-deprotection strategies for PPARγ activation. Competitive binding analysis of the synthesized molecules using time-resolved fluorescence resonance energy transfer (FRET) method was carried out, and the IC50 values were determined. The symmetrically substituted derivatives have shown greater binding affinity than unsymmetrically substituted derivatives. Nitrobenzyl and cyanophenyl substituted derivatives have shown reasonable binding affinities (10.1 and 6.5 µM, respectively), while mono and diacetate derivatives were found inactive. Molecular dynamics simulations show that the designed compounds have interaction profiles similar to partial agonists. The most significant finding of our study is that BA derivatives with symmetrically substituted weakly polar side chains result in the desired moderate level of PPARγ binding affinities.

Structural Rigidity and Protein Thermostability in Variants of Lipase A from Bacillus subtilis.

Understanding the origin of thermostability is of fundamental importance in protein biochemistry. Opposing views on increased or decreased structural rigidity of the folded state have been put forward in this context. They have been related to differences in the temporal resolution of experiments and computations that probe atomic mobility. Here, we find a significant (p = 0.004) and fair (R2 = 0.46) correlation between the structural rigidity of a well-characterized set of 16 mutants of lipase A from Bacillus subtilis (BsLipA) and their thermodynamic thermostability. We apply the rigidity theory-based Constraint Network Analysis (CNA) approach, analyzing directly and in a time-independent manner the statics of the BsLipA mutants. We carefully validate the CNA results on macroscopic and microscopic experimental observables and probe for their sensitivity with respect to input structures. Furthermore, we introduce a robust, local stability measure for predicting thermodynamic thermostability. Our results complement work that showed for pairs of homologous proteins that raising the structural stability is the most common way to obtain a higher thermostability. Furthermore, they demonstrate that related series of mutants with only a small number of mutations can be successfully analyzed by CNA, which suggests that CNA can be applied prospectively in rational protein design aimed at higher thermodynamic thermostability.

VisualCNA: a GUI for interactive constraint network analysis and protein engineering for improving thermostability.

UNLABELLED: Constraint network analysis (CNA) is a graph theory-based rigidity analysis approach for linking a biomolecule's structure, flexibility, (thermo)stability and function. Results from CNA are highly information-rich and require intuitive, synchronized and interactive visualization for a comprehensive analysis. We developed VisualCNA, an easy-to-use PyMOL plug-in that allows setup of CNA runs and analysis of CNA results linking plots with molecular graphics representations. From a practical viewpoint, the most striking feature of VisualCNA is that it facilitates interactive protein engineering aimed at improving thermostability. AVAILABILITY AND IMPLEMENTATION: VisualCNA and its dependencies (CNA and FIRST software) are available free of charge under GPL and academic licenses, respectively. VisualCNA and CNA are available at http://cpclab.uni-duesseldorf.de/software; FIRST is available at http://flexweb.asu.edu.

Quality matters: extension of clusters of residues with good hydrophobic contacts stabilize (hyper)thermophilic proteins.

Identifying determinant(s) of protein thermostability is key for rational and data-driven protein engineering. By analyzing more than 130 pairs of mesophilic/(hyper)thermophilic proteins, we identified the quality (residue-wise energy) of hydrophobic interactions as a key factor for protein thermostability. This distinguishes our study from previous ones that investigated predominantly structural determinants. Considering this key factor, we successfully discriminated between pairs of mesophilic/(hyper)thermophilic proteins (discrimination accuracy: ∼80%) and searched for structural weak spots in E. coli dihydrofolate reductase (classification accuracy: 70%).

CNA web server: rigidity theory-based thermal unfolding simulations of proteins for linking structure, (thermo-)stability, and function.

The Constraint Network Analysis (CNA) web server provides a user-friendly interface to the CNA approach developed in our laboratory for linking results from rigidity analyses to biologically relevant characteristics of a biomolecular structure. The CNA web server provides a refined modeling of thermal unfolding simulations that considers the temperature dependence of hydrophobic tethers and computes a set of global and local indices for quantifying biomacromolecular stability. From the global indices, phase transition points are identified where the structure switches from a rigid to a floppy state; these phase transition points can be related to a protein's (thermo-)stability. Structural weak spots (unfolding nuclei) are automatically identified, too; this knowledge can be exploited in data-driven protein engineering. The local indices are useful in linking flexibility and function and to understand the impact of ligand binding on protein flexibility. The CNA web server robustly handles small-molecule ligands in general. To overcome issues of sensitivity with respect to the input structure, the CNA web server allows performing two ensemble-based variants of thermal unfolding simulations. The web server output is provided as raw data, plots and/or Jmol representations. The CNA web server, accessible at http://cpclab.uni-duesseldorf.de/cna or http://www.cnanalysis.de, is free and open to all users with no login requirement.

Constraint Network Analysis (CNA): a Python software package for efficiently linking biomacromolecular structure, flexibility, (thermo-)stability, and function.

For deriving maximal advantage from information on biomacromolecular flexibility and rigidity, results from rigidity analyses must be linked to biologically relevant characteristics of a structure. Here, we describe the Python-based software package Constraint Network Analysis (CNA) developed for this task. CNA functions as a front- and backend to the graph-based rigidity analysis software FIRST. CNA goes beyond the mere identification of flexible and rigid regions in a biomacromolecule in that it (I) provides a refined modeling of thermal unfolding simulations that also considers the temperature-dependence of hydrophobic tethers, (II) allows performing rigidity analyses on ensembles of network topologies, either generated from structural ensembles or by using the concept of fuzzy noncovalent constraints, and (III) computes a set of global and local indices for quantifying biomacromolecular stability. This leads to more robust results from rigidity analyses and extends the application domain of rigidity analyses in that phase transition points ("melting points") and unfolding nuclei ("structural weak spots") are determined automatically. Furthermore, CNA robustly handles small-molecule ligands in general. Such advancements are important for applying rigidity analysis to data-driven protein engineering and for estimating the influence of ligand molecules on biomacromolecular stability. CNA maintains the efficiency of FIRST such that the analysis of a single protein structure takes a few seconds for systems of several hundred residues on a single core. These features make CNA an interesting tool for linking biomacromolecular structure, flexibility, (thermo-)stability, and function. CNA is available from http://cpclab.uni-duesseldorf.de/software for nonprofit organizations.