A pilot macromolecular 3D structure determination project - Year 10
Title: A pilot macromolecular 3D structure determination project - Year 10
DNr: SNIC 2022/5-414
Project Type: SNIC Medium Compute
Principal Investigator: Martin Moche <Martin.Moche@ki.se>
Affiliation: Linköpings universitet
Duration: 2022-09-01 – 2023-09-01
Classification: 10203 10601
Homepage: https://www.nsc.liu.se/support/presto/


In 2013, Protein Science Facility from Karolinska Institutet in Stockholm and National Supercomputer Centre (NSC) in Linköping started a pilot project to evaluate Macromolecular X-ray crystallography (MX) applications performance at NSC Triolith. In 2015, the Swedish light source MAX IV decided to fund an extension called PReSTO for MX, Nuclear Magnetic Resonance (NMR) and cryo-electron microscopy (cryo-EM). In 2018, the Swedish Research Council granted funds to this project (dnr. 2018-06479) Members of this project have access to the PReSTO installation via Swedish National Infrastructure of Computing (SNIC). Thinlinc software from Cendio supports the structural biology workflow by enabling remote graphic applications such as coot/chimera/ccp4mg/pymol for interactive model building and visualization of protein structure and surface properties. Since 2017, we use easybuild when installing PReSTO with several advantages to a standard HPC installation such as A) software environments can be send to compute nodes B) software build and runtime dependencies are made explicit in easyconfigs and easyblocks C) the version controlled software stack can be moved into new hardware with minor effort. PReSTO is now available at NSC Tetralith, LUNARC Aurora, the MAX IV online and offline clusters, and the local NSC clusters Sigma and Berzelius. The PReSTO homepage is designed for newcomers pointing towards structural biology software developer manuals, default option batch scripts and slurm configurations for a few GUIs with inherent slurm scheduler. We also developed a PReSTO menu that enable users to A) launch MX software at login or compute nodes where appropriate B) select compute node time and core number C) select output directory for some software such as hkl2map. Code optimization are made to adapt the forkxds script of popular XDS package to slurm, and XDSAPP3 to set local parameters according to slurm allocation. In 2020, MAX IV developed a fragment screening web application on top of the stable PReSTO installation(1) and PReSTO were acknowledged for making Serial X-ray crystallography software available to Swedish researchers(2) doing time resolved diffraction experiments using X-ray Free Electron Lasers (XFEL). During 2022 we transfer the PReSTO installation to LiU local NSC Berzelius cluster using easybuild. We had a first installation of Scipion, Relion, EMAN2, CisTEM, Gctf, Gautomatch, MotionCor2, and CTFFIND at Tetralith and added CryoSPARC when transfer to NSC Berzelius. Today users may collect cryo-EM data at SciLifeLab Umeå site and perform live data processing using “CryoSPARC live” at NSC Berzelius. There is a need to spread the word and know-how on how to do this across more users of SciLifeLab cryo-EM facilities. We added instructions to include grant numbers when acknowledging SNIC and PReSTO on our homepage. 1. G. M. A. Lima et al., FragMAXapp: crystallographic fragment-screening data-analysis and project-management system. ActaD 77, 799-808 (2021). 2. V. Srinivas et al., High-Resolution XFEL Structure of the Soluble Methane Monooxygenase Hydroxylase Complex with its Regulatory Component at Ambient Temperature in Two Oxidation States. JACS 142, 14249-14266 (2020).