BioExcel-2 D3.4 –User Community Support and Engagement Report (halftime update)

This report describes work carried out by the BioExcel Centre of Excellence to support and engage with its target user communities during the first 18 months of BioExcel-2, including work done to help ensure that BioExcel activities in all work packages - not only WP3 but also the software and workflow tool development (WP1 & WP2), training (WP4) and sustainability (WP5) - is user driven. During the first half of 2020, as well as executing COVID-19 related research work (see D3.3 - Use Case Progress Report), BioExcel has adapted to support efforts of the wider computational biomolecular research community to tackle the pandemic, including by facilitating research data sharing, support for BioExcel is meeting and exceeding stretch targets on all WP3 KPIs, having supported 600 community members and answered 880 questions on the AskBioExcel support forum and GROMACS mailing list during PM1-PM15, and during PM1-PM18 gathered input from 1200 community members and made available 23 tutorials to users. The Centre has delivered substantial amounts of support, especially to a well-established and ever growing user base corresponding to the core applications GROMACS, PMX and HADDOCK. The execution of free energy calculations by combining functionalities of GROMACS and PMX has been confirmed as an area of very strong interest from both academia and industry. Requests for support are accordingly high, and a pilot project with an industrial partner is in the process of being initiated. The high demand suggests opportunities exist for a BioExcel Legal Entity to sustain itself in part by focusing on meeting this demand. Steps are being taken to support the community in a scalable way by producing tutorials and best practice guides. Workflow tools also play a role, as the containerised execution of large numbers of free energy calculations orchestrated by the PyCOMPSs execution engine to efficiently make use of large-scale parallel HPC resources are enabling rapid massive screening of therapeutics on HPC platforms. As the Use Case work using these tools progresses, the Centre will be able to further leverage the capability demonstrating work done there to support the wider community make use of these tools to perform large-scale free energy calculations. In contrast to GROMACS, PMX and HADDOCK, there is still not a large user base actively using CP2K for biomolecular simulation using the hybrid QM/MM approach. BioExcel is providing live training sessions and creating, improving and updating tutorials that users can study by themselves to get started using CP2K for QM/MM simulation of biomolecular systems, and is in the process of producing a best practice guide. With regards to the GROMACS-CP2K interface, this is now at a maturity level where it can start to be used for Use Case work as planned in the original BioExcel-2 project timeline (see D3.3 - Use Case Progress Report and D1.4 - Project Release of Core Applications). CP2K is simultaneously being optimised for scaling and performance and the interface tested, with work underway to discover and document best practice for joint usage of GROMACS together with CP2K for performing QM/MM simulation including for Use Case demonstrator research work. In the second half of BioExcel-2, the Centre will draw on this experience to support and train the community. In addition to training plans, events are also being planned to help stimulate uptake of CP2K (+GROMACS) for biomolecular research using QM/MM simulation. This is being done by tackling both the hurdle of arriving at a scientifically valid, accuracy-controlled QM treatment of a user’s system using DFT, and issues surrounding usage of CP2K’s particular implementation of a given DFT approach and how to concretely use the software to obtain meaningful results in an efficient manner, with or without GROMACS. S urveys of GROMACS and QM/MM user communities have yielded insights that have already helped shape support activities such as the content of best practice guides, and they will continue to do so. Survey results are also feeding into development work, both by direct identification of feature and performance priorities and, in the case of CP2K, by ensuring that the QM/MM benchmarks used to profile and improve the code’s performance and parallel scaling include a range of system types, sizes, and model parameters that are representative of the priorities of the user community. Further survey work may also help identify additional opportunities for the Centre to facilitate uptake of workflow tools such as the BioExcel building blocks (bioBB) and the PyCOMPSs execution manager, which in a number of BioExcel Use Cases are key to enabling efficient exploitation of large-scale computing resources to demonstrate how to tackle grand scientific challenges.This is the public version of the deliverable where the section "Industry" has been redacted to retain confidentiality