Periodic Reporting for period 2 - BioExcel (Centre of Excellence for Biomolecular Research)
Periodo di rendicontazione: 2017-05-01 al 2018-12-31
The goal of the BioExcel project was to establish a Centre of Excellence for Biomolecular Research (BioExcel) that becomes a central hub for biomolecular modeling and simulations. The centre will enable better science by improving the most popular biomolecular software and spreading best practices and expertise among the communities through consultancy and training. Research and expertise in the centre covers structural and functional studies of the main building blocks of living organisms - proteins, DNA, saccharides, membranes, solvents and small molecules such as drug compounds.
The centre includes competence on dynamics, docking, screening and simulations of these building blocks, as well as analysis, refinement and classification of experimental structural data. Our expertise also covers techniques to model interactions between molecules, hybrid-QM/MM methods (targeting enzyme reactions) and coarse-level models (bridging applicability to higher levels of structural organization). It is contributing to important societal challenges in Health and Food and has wide impact in Academia as well as Industry, particularly the Pharma, Biotech, Chemical, and Food industries.
The specific goals of the centre are:
• Improve the performance, efficiency, and scalability of selected software packages of high importance for biomolecular Life Science research on next-generation HPC systems.
• Improve the usability of ICT technologies for biomolecular researchers, both in academia and industry. According to user needs, efficient workflow environments with associated data handling and integration of simulation and analysis codes will be devised.
• Promote best practices and train end users to make best use of both software and computational infrastructure.
• Develop appropriate governance structures and a sustainable business plan
The centre includes competence on dynamics, docking, screening and simulations of these building blocks, as well as analysis, refinement and classification of experimental structural data. Our expertise also covers techniques to model interactions between molecules, hybrid-QM/MM methods (targeting enzyme reactions) and coarse-level models (bridging applicability to higher levels of structural organization). It is contributing to important societal challenges in Health and Food and has wide impact in Academia as well as Industry, particularly the Pharma, Biotech, Chemical, and Food industries.
The specific goals of the centre are:
• Improve the performance, efficiency, and scalability of selected software packages of high importance for biomolecular Life Science research on next-generation HPC systems.
• Improve the usability of ICT technologies for biomolecular researchers, both in academia and industry. According to user needs, efficient workflow environments with associated data handling and integration of simulation and analysis codes will be devised.
• Promote best practices and train end users to make best use of both software and computational infrastructure.
• Develop appropriate governance structures and a sustainable business plan
Objective 1. Excellence in Biomolecular Software:
The final releases of the pilot applications provide much improved performance and scalability. GROMACS, in particular, comes with support for all current CPU and accelerator platforms, including expected future exascale systems, with more than double performance since the start of BioExcel; and our novel QM/MM interface of CPMD has shown to scale efficiently on all the Tier-0 resources, enabling QM/MM simulations with CPMD for PRACE projects. Significant efforts have also been spent on reducing the time-to-solution for common free energy and sampling simulations of small systems – identified as a priority from industrial and academic users – by enabling ensemble simulations that automatically optimize sampling. HADDOCK, which is an example of an HTC application has been expanded with options that will make it run much more efficiently in a massive, distributed exascale scenario.
Objective 2. Excellence in Usability:
We have developed a rich library of biomolecular building blocks covering the most common needs of researchers. Workflows, based on those blocks, have been successfully tested and launched for wider adoption. Web-based portals such as HADDOCK and PMX have been updated to provide excellent user experience. The role of BioExcel within ELIXIR has grown to become the main contributor of best practices in the domain of structural and computational biomolecular science. A key element of interoperability is the emerging Common Workflow Language, which BioExcel and ELIXIR have adapted early on and helped evolve as a community standard, while we are also supporting more interactive workflow environments like KNIME, Galaxy and Jupyter Notebook.
Objective 3. Competence-building among academia and industry:
The results have been disseminated widely to the community and industry. We were involved in the organisation of over 30 events. Highlights include impact through activities with large bio-infrastructures such as ELIXIR; countries from the EU 13 area (via VI-SEEM VRE); international foothold (via MolSSI, US, FDA/BioCompute); good relationships established with industry contacts from several pharma companies, including site-visits to Janssen and UCB. The webinar series have been very popular, some of them with over 1000 views. Our training program has been refined, and support material in the knowledgebase expanded. Our support forum, AskBioExcel, is being well-used: as a whole, the forum has had about 21,000 page views from logged-in users, about 210,000 page views, and the top posts on this forum have been viewed over 2000 times.
Objective 4. Work towards a sustainable structure:
The business plan was refined and finalized. We have identified several products and services (particularly workflows and training) suitable for commercialization. They have been analyzed in terms of the current market and competitor landscape, and further tested during a focus group meeting with pharma industry representatives. Various funding schemes and modalities were extensively studied. Viable governance and legal structures were identified. Registration of a legal entity is planned for 2019.
The final releases of the pilot applications provide much improved performance and scalability. GROMACS, in particular, comes with support for all current CPU and accelerator platforms, including expected future exascale systems, with more than double performance since the start of BioExcel; and our novel QM/MM interface of CPMD has shown to scale efficiently on all the Tier-0 resources, enabling QM/MM simulations with CPMD for PRACE projects. Significant efforts have also been spent on reducing the time-to-solution for common free energy and sampling simulations of small systems – identified as a priority from industrial and academic users – by enabling ensemble simulations that automatically optimize sampling. HADDOCK, which is an example of an HTC application has been expanded with options that will make it run much more efficiently in a massive, distributed exascale scenario.
Objective 2. Excellence in Usability:
We have developed a rich library of biomolecular building blocks covering the most common needs of researchers. Workflows, based on those blocks, have been successfully tested and launched for wider adoption. Web-based portals such as HADDOCK and PMX have been updated to provide excellent user experience. The role of BioExcel within ELIXIR has grown to become the main contributor of best practices in the domain of structural and computational biomolecular science. A key element of interoperability is the emerging Common Workflow Language, which BioExcel and ELIXIR have adapted early on and helped evolve as a community standard, while we are also supporting more interactive workflow environments like KNIME, Galaxy and Jupyter Notebook.
Objective 3. Competence-building among academia and industry:
The results have been disseminated widely to the community and industry. We were involved in the organisation of over 30 events. Highlights include impact through activities with large bio-infrastructures such as ELIXIR; countries from the EU 13 area (via VI-SEEM VRE); international foothold (via MolSSI, US, FDA/BioCompute); good relationships established with industry contacts from several pharma companies, including site-visits to Janssen and UCB. The webinar series have been very popular, some of them with over 1000 views. Our training program has been refined, and support material in the knowledgebase expanded. Our support forum, AskBioExcel, is being well-used: as a whole, the forum has had about 21,000 page views from logged-in users, about 210,000 page views, and the top posts on this forum have been viewed over 2000 times.
Objective 4. Work towards a sustainable structure:
The business plan was refined and finalized. We have identified several products and services (particularly workflows and training) suitable for commercialization. They have been analyzed in terms of the current market and competitor landscape, and further tested during a focus group meeting with pharma industry representatives. Various funding schemes and modalities were extensively studied. Viable governance and legal structures were identified. Registration of a legal entity is planned for 2019.
BioExcel acts as a bridge between e-Infrastructure providers, software developers, academic and industrial researchers within molecular life science, and other key stakeholders in computing. In so doing, BioExcel stimulates collaborative software development and use of e-Infrastructure for the benefit of European innovation and competitiveness, primarily via three core innovations:
Innovation 1: More efficient, scalable, and functional software
Impact: Through increasing the efficiency, scalability, and functionality of key biomolecular software, the computational costs will be drastically reduced providing a shorter time to solution, better exploitation of the available hardware, and thus reducing the development cycle and increasing the effectiveness of key products such as new drugs.
Results: We have delivered code releases with improved performance and scalability, new features introduced and support for cutting-edge hardware platforms.
Innovation 2: Improved access to data- and e-Infrastructures
Impact: Lowering the barrier to access data and e-Infrastructures will facilitate the uptake of computational technologies in the target sectors, improve their efficiencies and thus again directly contribute to increased efficiency and competitiveness.
Results: We have developed a rich library of biomolecular building blocks covering the most common needs of researchers. Workflows, based on those blocks, have been successfully tested and launched for wider adoption.
Innovation 3: Improved competence among academia and industry
Impact: By increasing the competence on using computational methods, providing support, and related training programs, BioExcel provides an innovative portfolio of services. It contributes to improved research output and fosters the international competitiveness of European academia & industry.
Results: We ran numerous topical workshops, hands-on and training events to address the gaps which were identified in the beginning of the project. Highlights include the impact through activities with large bio-infrastructures such as ELIXIR; countries from the EU 13 area (via VI-SEEM VRE); international foothold (via MolSSI, US); industrial support (site-visits to Janssen and UCB). The webinar series have been very popular, some with over 1000 views. Our training program has been refined, and support material in the knowledgebase expanded.
Innovation 1: More efficient, scalable, and functional software
Impact: Through increasing the efficiency, scalability, and functionality of key biomolecular software, the computational costs will be drastically reduced providing a shorter time to solution, better exploitation of the available hardware, and thus reducing the development cycle and increasing the effectiveness of key products such as new drugs.
Results: We have delivered code releases with improved performance and scalability, new features introduced and support for cutting-edge hardware platforms.
Innovation 2: Improved access to data- and e-Infrastructures
Impact: Lowering the barrier to access data and e-Infrastructures will facilitate the uptake of computational technologies in the target sectors, improve their efficiencies and thus again directly contribute to increased efficiency and competitiveness.
Results: We have developed a rich library of biomolecular building blocks covering the most common needs of researchers. Workflows, based on those blocks, have been successfully tested and launched for wider adoption.
Innovation 3: Improved competence among academia and industry
Impact: By increasing the competence on using computational methods, providing support, and related training programs, BioExcel provides an innovative portfolio of services. It contributes to improved research output and fosters the international competitiveness of European academia & industry.
Results: We ran numerous topical workshops, hands-on and training events to address the gaps which were identified in the beginning of the project. Highlights include the impact through activities with large bio-infrastructures such as ELIXIR; countries from the EU 13 area (via VI-SEEM VRE); international foothold (via MolSSI, US); industrial support (site-visits to Janssen and UCB). The webinar series have been very popular, some with over 1000 views. Our training program has been refined, and support material in the knowledgebase expanded.