PMDA - Parallel Molecular Dynamics Analysis

Abstract

MDAnalysis is an object-oriented Python library to analyze trajectories from molecular dynamics (MD) simulations in many popular formats. With the development of highly optimized MD software packages on high performance computing (HPC) resources, the size of simulation trajectories is growing up to many terabytes in size. However efficient usage of multicore architecture is a challenge for MDAnalysis, which does not yet provide a standard interface for parallel analysis. To address the challenge, we developed PMDA, a Python library that builds upon MDAnalysis to provide parallel analysis algorithms. PMDA parallelizes common analysis algorithms in MDAnalysis through a task-based approach with the Dask library. We implement a simple split-apply-combine scheme for parallel trajectory analysis. The trajectory is split into blocks, analysis is performed separately and in parallel on each block ("apply"), then results from each block are gathered and combined. PMDA allows one to perform parallel trajectory analysis with pre-defined analysis tasks. In addition, it provides a common interface that makes it easy to create user-defined parallel analysis modules. PMDA supports all schedulers in Dask, and one can run analysis in a distributed fashion on HPC machines, ad-hoc clusters, a single multi-core workstation or a laptop. We tested the performance of PMDA on single node and multiple nodes on a national supercomputer. The results show that parallelization improves the performance of trajectory analysis and, depending on the analysis task, can cut down time to solution from hours to minutes. Although still in alpha stage, it is already used on resources ranging from multi-core laptops to XSEDE supercomputers to speed up analysis of molecular dynamics trajectories. PMDA is available as open source under the GNU General Public License, version 2 and can be easily installed via the pip and conda package managers.

Zinc Dependent Conformational Changes in the Cation Diffusion Facilitator YiiP from S. oneidensis

“…RMSD's of backbone atoms, all binding site residues and ligand atoms were calculated to assess the stability of the Zn-bound structures. Radial distribution functions between the zinc ion and binding site ligand atoms were compared with the distances from crystal structures using RDF.InterRDFs in Parallel MD Analysis (PMDA) (Fan et al, 2019). Development of antibody fragments that recognize YiiP.…”

Parallel performance of molecular dynamics trajectory analysis

“…In other studies, better scaling was observed for analysis tasks that were more compute‐intensive than the RMSD calculation, such as a radial distribution function calculation, 26,52 that is, analysis tasks that could be characterized as compute‐bound. Such behavior is expected, as the contribution from the parallel part of the program that requires neither I/O nor communication is increased.…”

“…Although straggler tasks are a common challenge arising in parallel analysis and are well known in the data analysis community (see Section 2.1), there is, to our knowledge, little discussion about this problem in the biomolecular simulation community. Our own experience with a MapReduce approach in MDAnalysis 24,26 suggested that stragglers might be a somewhat under‐appreciated problem. Therefore, in the present work we wanted to better understand requirements for efficient parallel analysis of MD trajectories in MDAnalysis , but to also provide more general guidance that could benefit developments in other libraries.…”

“…Previously, we had used a parallel split‐apply‐combine approach 23 to study the performance of the commonly performed “RMSD fitting” analysis problem, 24-26 which calculates the minimal root mean squared distance (RMSD) of the positions of a subset of atoms to a reference conformation under optimization of rigid body translations and rotations 7,27,28 . We had investigated two parallel implementations, one using Dask 29 and one using the message passing interface (MPI) with mpi4py 30,31 …”

LiPyphilic: A Python toolkit for the analysis of lipid membrane simulations

“…In this respect, we are particularly interested in the development of Parallel MDAnalysis (PMDA). 66 PMDA is based on MDAnalysis and uses Dask to parallelise the analysis modules. 67 The analysis classes inherit a modified abstract base class that is specifically designed to make the parallelisation straightforward.…”