Post-workshop update – the best practices covered in this workshop have been captured in the form of a guide for future reference:

Best Practice Guide for QM/MM Simulation of Biomolecular Systems

 

Context

Users of hybrid QM/MM – quantum mechanics / molecular mechanics – approaches for biomolecular simulation face two key challenges.

The first challenge revolves around choosing an adequate quantum treatment for their biomolecular system of interest that is based on a reasoned

understanding of the accuracy and limitations of the chosen computational approach in light of what is known about the underlying quantum chemistry. This includes the QM theory level, for example the choice of density functional in the case of a DFT treatment, the size of the QM region, etc.  Many inexperienced practitioners use default or common choices and hope for the best, or refer to previously published work as justification for adequacy whilst treating a system with substantially distinct underlying quantum chemistry. In a recent survey carried out by BioExcel amongst the computational biomolecular research community, half of respondents reported not knowing how to choose suitable QM parameters to have significantly hindered their usage of QM/MM.

The second challenge is to choose a suitable reaction coordinate (and with it a suitable method to sample/minimize the system along that coordinate) in the high-dimensional configuration space of a typical (bio)chemical system. Reducing the space to a chemically meaningful set of reaction coordinates again requires a deep understanding of the underlying chemistry.

The discussions that really matter with regards to choosing the QM approach for QM/MM simulation of biomolecular systems and/or defining suitable reaction coordinates are ones that have been ongoing in the computational chemistry community for decades. These difficulties hinder the successful exploitation of QM/MM simulation, and are a prior consideration to what particular software package or combination of packages is used.

Goal

The overall goal of the workshop is to enable researchers with substantial experience and success using QM/MM for biomolecular simulation to share best practices regarding its effective application, both in general and in relation to their specific area of study. Aims are to draw attention not only to fundamental underlying theoretical issues but also to tricky critical practical aspects that are not necessarily described in the literature, and to highlight pitfalls as a warning to less experienced practitioners. As well as benefiting less experienced practitioners in their use of QM/MM, the workshop aims to enable experienced users to exchange perspectives grounded in the study of a variety of areas and to thereby reflect on common best practices as well as outstanding challenges.

Target audience

The workshop is aimed at anyone in the computational biomolecular research community with an interest in QM/MM, regardless of their current level of experience actively using QM/MM.

Format

The workshop takes place virtually and consists of a number of webinars by QM/MM expert invited speakers delivered over a period of time between October 30th 2020 and the end of January 2021, followed by a livestreamed online panel discussion and audience Q&A between the speakers. Recordings of the webinars will be available on the BioExcel YouTube channel.

Speakers

Maria João Ramos Adrian Mulholland Maria Khrenova
Ulf Ryde Carme Rovira Janez Mavri

Gerrit Groenhof
(co-organiser)

Schedule

Friday October 30th 2020:  Gerrit Groenhof – Workshop Kick-off webinar   (recording, slides)

Friday November 13th 2020:  Maria Khrenova – Validation of DFT functionals in QM/MM simulations of enzymatic reactions initiated by the nucleophilic attack   (recording, slides)

Friday November 26th 2020 Ulf Ryde – Obtaining accurate structures and energies with QM/MM  (recording, slides

Thursday December 10th 2020:  Maria João Ramos – Studies on enzyme-catalysed reactions  (recording, slides)

Monday December 14th 2020:  Adrian Mulholland – Towards chemical accuracy in QM/MM modelling of enzyme catalytic mechanisms and protein-ligand binding (recording, slides)

Friday January 8th 2021:  Janez Mavri – Multiscale Simulation of Monoaminergic System: Applications to Neurodegeneration (recording, slides)

Tuesday January 12th 2021:  Carme Rovira – Modeling catalytic mechanisms in carbohydrate-active enzymes with QM/MM MD methods (recording, slides)

Friday January 29th: Panel Session  (recording, slides)