Over 75% of disease-involved proteins cannot be readily targeted by conventional chemical biology approaches. New approaches are needed to increase the scope of molecular medicine. Cryptic binding pockets, i.e. pockets that transiently form in a folded protein, but are not apparent in the crystal structure of the unliganded apo-form, offer outstanding opportunities to target proteins otherwise deemed 'undruggable' and are thus of considerable interest in academia and the pharmaceutical industry. Unfortunately, not only they are notoriously difficult to identify, but also the molecular mechanism by which they form is still debated. The aim of this collaborative project is to address the knowledge gaps and develop an efficient computational platform based on atomistic molecular simulations to systematically detect druggable cryptic pockets in targets of biopharmaceutical interest. The platform will build on our successful experience in developing and applying enhanced-sampling simulation algorithms to molecular recognition, and will be extensively tested on validated drug targets harbouring cryptic sites. The computational results will be further validated on novel targets by a combination of experiments in collaboration with an industrial partner (UCB).

Atomistic simulations are the main application of high-performance computing research, and increasingly underpin innovative R&D processes in the chemical and life sciences industry. OpenMM is the fastest growing atomistic simulation engine among the current ecosystem of open-source academic software. Originally targeting a biomolecular simulation audience, the OpenMM user base is growing exponentially and has permeated diverse related domains, including materials modelling, quantum chemistry, structural bioinformatics, chemoinformatics, artificial intelligence and machine learning. The success of OpenMM is down to a design that achieves an excellent tradeoff between extensibility (via a robust user interface) and performance on GPUs (via auto generated CUDA kernels) for molecular dynamics (MD) simulations. OpenMM is used standalone or via plugins to other atomistic simulation engines, providing access to GPU-accelerated MD simulation capabilities for the whole atomistic simulation ecosystem. We have surveyed the OpenMM user community to identify its most pressing needs. OpenMM is currently maintained by a single core developer who can no longer support the training and support needs of its rapidly growing user community. We will transition OpenMM to a more sustainable community-driven development model. We will develop training resources to upskill users, and engage the community to widen participation in developing and maintaining OpenMM functionality. Machine learning (ML) potentials have the potential to revolutionise the future of atomistic simulation methodologies. Our community survey has identified strong interest in ANI neural network and GAP Gaussian process regression methods. We will deliver a self-contained GPU-optimised GAP implementation in OpenMM and coordinate with project partners working on an OpenMM ANI implementation to offer the community a library of ML potentials that can be readily plugged into existing simulation engines. OpenMM must adapt to scientific (ML potentials) and technological (increased hardware heterogeneity) drivers to continue offering its user base an optimised tradeoff between speed and ease of modification over the coming decade. We will integrate in OpenMM a multiple level intermediate representation compiler (MLIR) to auto generate from user-specified Python instructions optimised low-level code targeting diverse hardware. By enabling users to specify custom atomic featurisation techniques as OpenMM operations, which can be finely interleaved with Tensorflow or Pytorch operations, we will position OpenMM as the simulation engine of choice to support deployment of next generation ML potentials onto current GPUs and emerging AI-hardware accelerators. Our community has also required support to facilitate the combined use of independently developed OpenMM software solutions with other software from the broader atomistic simulation ecosystem. This research will develop a standardised interface to integrate OpenMM community software with CCPBioSim's interoperable Python framework BioSimSpace. We will demonstrate integration of all the work packages of this research via production of GAP ML pipelines for two use cases that target grand challenges in soft-condensed matter modeling (organocatalysis - recently recognised by the 2021 Nobel Prize in Chemistry- and protein-ligand binding). Altogether this research will position the OpenMM user community at the forefront of next-generation hybrid machine learning/molecular mechanics potentials for soft-condensed matter modelling. Deeper integrations with AI and HPC communities will pave the way for atomistic simulations to harness emerging exascale opportunities. Transitioning from a single developer to a community-driven development governance model will improve sustainability of the codebase and encourage greater adoption of OpenMM in associated academic communities and industry.

First-principles quantum-mechanical simulations based on density-functional theory (DFT), are today used hand in hand with experiment to design new materials. Conventional DFT has a computational effort which increases with the cube of the number of atoms and this limits the practical size of calculations. ONETEP is a world-leading UK-developed software package which uses a linear-scaling framework to enable calculations on much larger scales, uniquely without loss of accuracy compared to traditional methods. Thus ONETEP offers unmatched capabilities for constructing and simulating more realistic models of materials and including their environment in multiscale simulations. ONETEP has been developed from the beginning to take advantage of supercomputers. Due to its non-trivial formulation and wide-ranging functionality, it is a highly complex code consisting of around half a million lines of code. ONETEP is an academic community code which emerged from CCP9, the Collaborative Computational Project for the electronic structure of condensed matter, bringing together academics across disciplines, and forming the UK branch of the European Psi-k Network. In 2016 ONETEP became the flagship project of CCP9 and free to UK academics. Industrial exposure to ONETEP has resulted from close collaboration with BIOVIA, which has enabled integration with their Materials Studio user interface. This has led to considerable commercial impact and new industrial collaborations. Beyond the UK, ONETEP is gaining in popularity with developers in Ireland and China and users in many countries in Europe as well as the USA, China, Mexico and South Africa. As with all software, ONETEP needs to be continuously evolved and updated in order to stay at the cutting edge. This is particularly challenging for a large collaborative academic project that has evolved over two decades. Furthermore, a range of developments, such as excited states, electrochemistry, embedding and wavefunction methods, have required pervasive changes. Since they affect the core algorithms of the code, these changes have inevitably led to increased complexity. Thus the code now needs to adopt a new structure to ensure its continued growth. At the same time it is important to maintain and further widen the community of users and developers to fulfill its primary objective to cater for the needs of the scientific community. This project is targeted towards these two interconnected aims. It will re-engineer the code in its entirety, rationalising internal structure to allow further development and enhance the interoperability of existing functionality. Modern software engineering principles will be followed throughout, in close collaboration with the computational physics and chemistry groups of STFC SCD and research software engineers in Southampton, Warwick and Imperial. At the same time developments of new functionality to enable large-scale calculations of crystalline and semicrystalline materials will satisfy a demand in this area by many researchers, such as in the CCP9 and the solid state microscopy and spectroscopy communities at STFC Facilities. Workflow tools and coupling with the ChemShell QM/MM code will be developed to allow adoption of the code by the biomolecular simulations community. The code will also be ported to emerging supercomputing architectures with GPU accelerators. Thus the project will support the rapidly-expanding communities within solid-state materials and biochemistry that deploy first-principles quantum simulations based on DFT. The project will deliver significant communication, engagement, and expert training and mentoring of new users to overcome initial barriers to access and enable them to use the code to make impact in their diverse research areas. Training events for both users and developers of the code will be embedded within each community.

Nitrogen compounds play a crucial role in the earth's ecosystems, being continually converted from one form to another as they pass from the atmosphere to living organisms on land and in the sea. Nitric oxide gas (NO), for example, is a key intermediate in the global nitrogen cycle, and plays important roles in many processes in almost all forms of life, often acting as a signalling molecule. However, emissions of NO and the toxic gas nitrogen dioxide (collectively known as NOx) from heavy industry and motor vehicles alter the composition of nitrogen compounds in the atmosphere and are highly damaging both directly and indirectly to the human respiratory system. The removal of NOx from exhaust emissions is a pressing environmental concern and an important target for industrial catalysis research, an area of extreme importance to the UK economy. We propose to study the chemistry of nitrogen oxides in biological and industrial environments where a full understanding of how the gases are controlled is crucial but still lacking. In both cases the chemistry is controlled by transition metals: cytochrome c' proteins have evolved an extraordinary degree of control of NO through binding to an iron complex which discriminates against other diatomic gases, while in zeolite catalysts (microporous aluminosilicate structures) NOx gases can be converted into safer by-products at copper centres through the addition of ammonia in a process known as selective catalytic reduction (SCR). The precise mechanisms, however, are not currently proven. We will investigate the chemistry of nitrogen dioxide and nitrogen oxide in both systems by computational simulations performed on high performance clusters. The resulting data will be used to model spectroscopic signatures, i.e. how electromagnetic radiation (such as light or X-rays) interacts with matter. These will be compared with the results of infrared, Raman, UV-visible and X-ray absorption experiments on the two systems to better understand the processes involved in the chemical reactions, which will inform the future design of improved zeolite catalysts and bioengineered proteins. We will use quantum mechanical/molecular mechanical (QM/MM) modelling to identify the reaction mechanisms and calculate spectroscopic signatures of the two systems. In this approach the zeolite and protein active sites will be treated using a highly accurate, but computationally expensive, quantum mechanical level of theory, embedded in an environment described by an efficient classical calculation. New QM/MM methods will be implemented that can enable larger QM regions to be calculated and more accurate spectroscopic signatures including anharmonic vibrational effects. Importantly, our approach for combining computational modelling with experimental results will be generally applicable to any chemical processes in complex systems, including other industrial catalysts and biomolecules.

Molecular modelling has established itself as a powerful predictive tool for a large range of materials and phenomena whose intrinsic multiscale nature requires modelling tools able to capture their chemical, morphological and structural complexity. In the UK, the molecular modelling community, supported by the software, training and networking activities coordinated by the CCP5, has become, over the past 40 years, international-leading in this field. Building upon these successes, the new CCP5++ network will revolutionise the field of materials molecular modelling creating a new integrated community of modellers, experimentalists and data scientists that together will identify the new frontiers of the field and will transform the way these disciplines work together. To achieve its mission, the CCP5++ will coordinate and support an ambitious plan of meetings, sandpits, coding workshops, secondments and visitor schemes to cater for the large community of modellers, experimentalists and data scientists working on advanced materials. This support has proved to be vital to enable the UK condensed matter community to attain and maintain an international position at the forefront of such an intensely competitive field and will enable the UK researchers to identify and tackle major world challenges in-silico materials discovery. From the start the network memberships include key representatives of the experimental and data science communities, international software and modelling institutes, industrial collaborators and national HPC consortia and CCPs, that working together will shape the future of materials molecular modelling.