Search

This document is a final update to deliverables D3.4 - User Community Support and Engagement Report (halftime update) (PM18) and D3.5 - Best Practice Guides (PM24). It describes BioExcel’s community support and engagement activities from PM19 (July 2020) up until and including PM42 (June 2022). It includes a summary for this period of the community events BioExcel (co-)organised or participated in, the in-depth support we have offered to users, our work on standards and best practice, including updates to our Best Practice Guides, and how we have analysed user & community needs to drive our activities. During this reporting period BioExcel has: (Co-)organised and/or participated in 136 community events, attended by around 10 000 community members, at which we shared scientific results from the BioExcel Use Cases and/or the capabilities of BioExcel software, at external workshops and conferences and at training courses run together with PRACE, the EuroHPC National Competence Centres, and other HPC CoEs. We organised a workshop - “Best Practices in QM/MM Simulation of Biomolecular Systems”, which was attended by around 1000 people. Through our support and engagement work we have started seeing increased uptake of the new GROMACS-CP2K interface we developed, and we deem to have successfully seeded a new user community. Hosted 23 BioExcel Webinars, attended live by over 1700 people. Seen our YouTube channel gain significant momentum, with ~4x greater content consumption rate and ~3x faster growth in subscribers than during the previous reporting period. Provided direct in-depth support to over 1900 community members, including by answering around 1500 questions on the AskBioExcel and GROMACS forums combined. Created, improved, and made available 16 tutorials for users of BioExcel core software. Engaged in existing and new pilot projects with industry and pursued our planned strategy regarding what BioExcel will deliver for industrial users, namely: Prepare and establish BioExcel Enterprises Consolidate existing and establish new business opportunities Further define and validate the BioExcel value proposition Supported the National Centre for Supercomputing Applications in Bulgaria, which hosts the EuroHPC Discoverer system by providing lectures and hands-on training in BioExcel software functionality, methodologies and performance on two separate courses. Continued to participate in community development of standards such as the Common Workflow Language, FAIR computational workflows, FAIR Digital Objects, BioCompute Objects and Research Object Crate Expanded the Best Practice Guides (BPGs) first published at PM24 as part of deliverable D3.5 - Best Practice Guides as we described and planned in that deliverable. We have: Updated the CP2K biomolecular QM/MM BPG to include usage of GROMACS together with CP2K for QM/MM simulation of biomolecular systems Captured the outcomes of our workshop on Best Practices in QM/MM Simulation of Biomolecular Systems in a new software-agnostic Best Practice Guide on general QM/MM modelling & simulation methodology. Added guidance to the GROMACS BPG detailing how to obtain good performance on two additional European HPC systems - JUWELS BOOSTER (PRACE) and Discoverer (EuroHPC) - and on using GROMACS postprocessing tools. Adapted activities on an ongoing basis, including after gathering input, including feedback, from over 1700 community members and analysing user needs. Interviewed 8 Key Opinion Leaders (KOLs) in a range of biomolecular modelling & simulation research areas to help us assess what we have achieved in BioExcel-2 and, crucially, to help us look ahead and calibrate our vision towards the exascale future. Designed and begun to prototype an Ambassador programme in order to strengthen links with local computational biomolecular research communities across the EU, starting with EU13 member states. &nbsp

This deliverable presents the final release of demonstration workflows for the BioExcel-2 project, including a wrap-up of all the work done for the BioBB library, its associated workflows, and a final analysis of the milestones achieved from the initial roadmap. The current release of the BioExcel Building Blocks library is presented in the first section, underlining the FAIR development process followed, including the automatic generation of the BioBB adapters from the Python documentation. The current available modules (categories) in the library are listed. Finally, the landing website and its different sections are briefly described. The second section enumerates the final collection of demonstration workflows, with a brief description of what each workflow does, the tools used, and links to the source code and documentation. The library versatility, validated with the collection of demonstration workflows, is explained, and links to the different workflow flavours are presented in a summary table. The power of the pre-exascale demonstration workflows is presented in the fourth section, with examples of workflows successfully applied to real scientific projects (success stories). The following section focuses on the broad range of possibilities to find and execute BioExcel workflows, from local or HPC installations, to web-based interfaces. Visibility and uptake is discussed in the sixth section, listing a number of training events, webinars and scientific drafts with the presence of the library and the demonstration workflows. The final section revisits the proposed roadmap for the WP2, analysing the status of the intended milestones at the end of the project, with comments on the progress, potential issues and redefined tasks.

Mutations in the kinase domain of the Epidermal Growth Factor Receptor (EGFR) can be drivers of cancer and also trigger drug resistance in patients under chemotherapy treatment based on kinase inhibitors use. A priori knowledge of the impact of EGFR variants on drug sensitivity would help to optimize chemotherapy and to design new drugs effective against resistant variants. To this end, we have explored a variety of in silico methods, from sequence-based to ‘state-of-the-art’ atomistic simulations. We did not find any sequence signal that can provide clues on when a drug-related mutation appears and what will be the impact in drug activity. Low-level simulation methods provide limited qualitative information on regions where mutations are likely to produce alterations in drug activity and can predict around 70% of the impact of mutations on drug efficiency. High-level simulations based on non-equilibrium alchemical free energy calculations show predictive power. The integration of these ‘state-of-the-art’ methods in a workflow implementing an interface for parallel distribution of the calculations allows its automatic and high-throughput use, even for researchers with moderate experience in molecular simulations.

14 pages, 6 figures, 1 table, supporting information https://pubs.acs.org/doi/10.1021/acs.jcim.2c01344.-- The code for this work is available at https://github.com/bioexcel/biobb_hpc_workflows/tree/condapack. MD simulations are available from our MDposit repository: https://mdposit-dev.bsc.es/#/browse?search=egfr, Mutations in the kinase domain of the epidermal growth factor receptor (EGFR) can be drivers of cancer and also trigger drug resistance in patients receiving chemotherapy treatment based on kinase inhibitors. A priori knowledge of the impact of EGFR variants on drug sensitivity would help to optimize chemotherapy and design new drugs that are effective against resistant variants before they emerge in clinical trials. To this end, we explored a variety of in silico methods, from sequence-based to “state-of-the-art” atomistic simulations. We did not find any sequence signal that can provide clues on when a drug-related mutation appears or the impact of such mutations on drug activity. Low-level simulation methods provide limited qualitative information on regions where mutations are likely to cause alterations in drug activity, and they can predict around 70% of the impact of mutations on drug efficiency. High-level simulations based on nonequilibrium alchemical free energy calculations show predictive power. The integration of these “state-of-the-art” methods into a workflow implementing an interface for parallel distribution of the calculations allows its automatic and high-throughput use, even for researchers with moderate experience in molecular simulations, This work has been supported by the BioExcel-2: Centre of Excellence for Computational Biomolecular Research (823830), the Spanish Ministry of Science (RTI2018-096704-B-100 and PID2020-116620GB-I00), and the Instituto de Salud Carlos III–Instituto Nacional de Bioinformática (ISCIII PT 17/0009/0007 cofunded by the Fondo Europeo de Desarrollo Regional). Funding was also provided by the MINECO Severo Ochoa Award of Excellence from the Government of Spain (awarded to IRB Barcelona). M.O. is an ICREA (Institució Catalana de Recerca i Estudis Avancats) Academia researcher. Nostrum Biodiscovery is supported by the Fundación Marcelino Botín (Mind the Gap), CDTI (Neotec grant EXP 00094141/SNEO-20161127), and a Torres Quevedo grant (PTQ2018-009992), With the institutional support of the ‘Severo Ochoa Centre of Excellence’ accreditation (CEX2019-000928-S)

Computer simulation techniques are gaining a central role in molecular pharmacology. Due to several factors, including the significant improvements of traditional molecular modelling, the irruption of machine learning methods, the massive data generation, or the unlimited computational resources through cloud computing, the future of pharmacology seems to go hand in hand with in silico predictions. In this review, we summarize our recent efforts in such a direction, centered on the unconventional Monte Carlo PELE software and on its coupling with machine learning techniques. We also provide new data on combining two recent new techniques, aquaPELE capable of exhaustive water sampling and fragPELE, for fragment growing.

In this study, we present a fully automatic platform based on our Monte Carlo algorithm, the Protein Energy Landscape Exploration method (PELE), for the estimation of absolute protein-ligand binding free energies, one of the most significant challenges in computer aided drug design. Based on a ligand pathway approach, an initial short enhanced sampling simulation is performed to identify reasonable starting positions for more extended sampling. This stepwise approach allows for a significant faster convergence of the free energy estimation using the Markov State Model (MSM) technique. PELE-MSM was applied on four diverse protein and ligand systems, successfully ranking compounds for two systems. Based on the results, current limitations and challenges with physics-based methods in computational structural biology are discussed. Overall, PELE-MSM constitutes a promising step toward computing absolute binding free energies and in their application into drug discovery projects.

12 páginas.- 6 figuras.- 1 tabla.- 84 referencias.- The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acscatal.0c03165, Epoxidation of unsaturated fatty acids by unspecific peroxygenases (UPOs) of the best-known long-UPO subfamily, including the Agrocybe aegerita (AaeUPO) and Coprinopsis cinerea enzymes, is reported here. To understand the different oxygenation patterns by members of the long-UPO and short-UPO subfamilies, the latter represented by the Marasmius rotula enzyme (MroUPO), fatty-acid diffusion into their heme pockets was simulated with the adaptive PELE software. Computational results shed light on the inability of AaeUPO to epoxidize oleic acid (C18:1), opposed to MroUPO, due to steric hindrances to harbor (with a good interaction energy) the substrate with the Δ9 C10 atom at a catalytically relevant distance (99%) formation of cis,cis-15,16-epoxyoctadeca-9,12-dienoic acid. The nearly complete conversion of α-linolenic acid by the two enzymes was transferred to a small preparative scale, the yield of purified product was estimated, its chemical structure analyzed by NMR, and more interestingly, stereoselective production of the 15(R),16(S) enantiomer (80-83% ee) assessed by chiral HPLC. This enzymatic synthesis overcomes the unspecificity of chemical epoxidation where the reaction cannot be restricted to the formation of monoepoxides as found during m-perchlorobenzoic acid oxidation of α-linolenic acid. Moreover, the variant was able to produce the unsaturated dicarboxylic fatty acid, together with subterminal oxygenation products, during partial conversion of oleic acid. These two noteworthy reactions had not been reported for any UPO described to date. © 2020 American Chemical Society. All rights reserved., This work has received funding from the Bio Based Industries Joint Undertaking under the European Union’s Horizon 2020 research and innovation programme under Grant Agreement No 792063 (“Development and pilot production of sustainable binder systems for wood based panels”, https://susbind.eu ), the CTQ2016-79138-R and BIO2017-86559-R projects of Spanish MINECO, the Secretaria d’Universitats i Recerca of Generalitat de Catalunya, and the European Social Fund (ESF-2019-FI-B-00154). The authors thank Novozymes A/S for supplying r CciUPO. MM acknowledges a Catalan Government doctoral grant.

Water is frequently found inside proteins, carrying out important roles in catalytic reactions or molecular recognition tasks. Therefore, computational models that aim to study protein–ligand interactions usually have to include water effects through explicit or implicit approaches to obtain reliable results. While full explicit models might be too computationally daunting for some applications, implicit models are normally faster but omit some of the most important contributions of water. This is the case of our in-house software, called protein energy landscape exploration (PELE), which uses implicit models to speed up conformational explorations as much as possible; the lack of explicit water sampling, however, limits its model. In this work, we confront this problem with the development of aquaPELE. It is a new algorithm that extends the exploration capabilities while keeping efficiency as it employs a mixed implicit/explicit approach to also take into account the effects of buried water molecules. With an additional Monte Carlo (MC) routine, a set of explicit water molecules is perturbed inside protein cavities and their effects are dynamically adjusted to the current state of the system. As a result, this implementation can be used to predict the principal hydration sites or the rearrangement and displacement of conserved water molecules upon the binding of a ligand. We benchmarked this new tool focusing on estimating ligand binding modes and hydration sites in cavities with important interfacial water molecules, according to crystallographic structures. Results suggest that aquaPELE sets a fast and reliable alternative for molecular recognition studies in systems with a strong water-dependency. This work has been supported by a predoctoral fellowship from the Government of Catalonia (2019FI_B_00154 to MM).

8 páginas.-- 4 figuras.-- 4 tablas.-- 43 referencias.-- Supporting information http://www.rsc.org/suppdata/c8/cy/c8cy02114g/c8cy02114g1.pdf, The recently discovered unspecific peroxygenases (UPOs) from the ascomycetes Chaetomium globosum and Humicola insolens were capable of selectively hydroxylating isophorone to 4-hydroxyisophorone (4HIP) and 4-ketoisophorone (4KIP), which are substrates of interest for the pharmaceutical and flavor-and-fragrance sectors. The model UPO from the basidiomycete Agrocybe aegerita was less regioselective, forming 7-hydroxyisophorone (and 7-formylisophorone) in addition to 4HIP. However, it was the most stereoselective UPO yielding the S-enantiomer of 4HIP with 88% ee. Moreover, using H. insolens UPO full kinetic resolution of racemic HIP was obtained within only 15 min, with >75% recovery of the R-enantiomer. Surprisingly, the UPOs from two other basidiomycetes, Marasmius rotula and Coprinopsis cinerea, failed to transform isophorone. The different UPO selectivities were rationalized by computational simulations, in which isophorone and 4HIP were diffused into the enzymes using the adaptive PELE software, and the distances from heme-bound oxygen in H2O2-activated enzyme to different substrate atoms, and the corresponding binding energies were analyzed. Interestingly, for process upscaling, full conversion of 10 mM isophorone was achieved with H. insolens UPO within nine hours, with total turnover numbers up to 5500. These biocatalysts, which only require H2O2 for activation, may represent a novel, simple and environmentally-friendly route for the production of isophorone derivatives., This work was supported by the EnzOx2 (H2020-BBI-PPP-2015-2-1-720297) EU-project, the AGL2014-53730-R (BIORENZYMERY) and CTQ2016-79138-R pojects of the Spanish MINECO (co-financed by FEDER) and the CSIC (201740E071) project. Novozymes (Bagsvaerd, Denmark) is acknowledged for providing samples of rCciUPO and rHinUPO.