MD 26 - Python Analysis Environment

Cluster: XLence (UNIMI Dipartimento di Scienze Farmacologiche e Biomolecolari) Module: md/26

---

Overview

The md/26 module is the Python analysis environment for molecular dynamics: MDAnalysis, MDTraj, RDKit, ProDy, ParmEd, JupyterLab and more, on Python 3.12.

Since 2026-09-30 the MD programs are separate modules, built without md26, and are no longer part of md/26:

Program	Module	Together with md/26
GROMACS 2025.0 + PLUMED 2.10.1	gromacs/2025.0	yes: module load md/26 gromacs/2025.0
PLUMED 2.10.1	plumed	yes: module load md/26 plumed
AMBER 26	amber/26	no (see below)

---

Loading

module load md/26

# Python analysis together with GROMACS (and PLUMED, which gromacs loads)
module load md/26 gromacs/2025.0

md/26 and amber/26 cannot be loaded together: Amber's Python modules are built for Amber's own Python (numpy 1.26) and would shadow the md26 ones. In a job that needs both, switch between steps: module unload md/26; module load amber/26.

MPI: mpi4py uses the system OpenMPI 4.1.6, the same as gmx_mpi and the system mpirun.

---

GROMACS Python API (gmxapi)

gmxapi is built against gromacs/2025.0: load both modules.

import gmxapi as gmx

md = gmx.mdrun(input=gmx.read_tpr('topol.tpr'))
md.run()

---

Python Analysis Stack

Trajectory Analysis

import MDAnalysis as mda

u = mda.Universe('system.prmtop', 'traj.nc')
protein = u.select_atoms('protein')

import mdtraj as md

traj = md.load('traj.dcd', top='system.pdb')
rmsf = md.rmsf(traj, traj, 0)

Cheminformatics and Structure

from rdkit import Chem
from rdkit.Chem import AllChem

mol = Chem.MolFromSmiles('CCO')
AllChem.EmbedMolecule(mol)

from prody import parsePDB, calcANM

atoms = parsePDB('protein.pdb')
anm = calcANM(atoms.select('calpha'))

Visualization in Jupyter

import nglview as nv
import MDAnalysis as mda

u = mda.Universe('system.prmtop', 'traj.nc')
view = nv.show_mdanalysis(u)
view

---

Available Python Packages

MD Analysis

Package	Description
MDAnalysis	Universal trajectory analysis
MDTraj	Fast trajectory analysis
pyedr	GROMACS .edr file reader
griddataformats	Volumetric grids (DX, CCP4)
nglview	Interactive molecular visualization (Jupyter)

Cheminformatics and Structure

Package	Description
RDKit	Cheminformatics
OpenBabel	Chemical file format conversion
Biopython	Bioinformatics
ProDy	Protein dynamics, NMA, ENM
ParmEd	Topology and parameter editing
propka	pKa prediction
ASE	Atomic Simulation Environment
geometric	Geometry optimizer (QM/MM)
cclib	QC output file parser

Statistics and ML

Package	Description
scikit-learn	Machine learning
scipy	Scientific computing
statsmodels	Statistical models
pingouin	Advanced statistics
numba	JIT compilation for fast analysis
umap-learn	UMAP dimensionality reduction
scanpy	Single-cell analysis
anndata	Annotated data (omics)

Plotting

Package	Description
matplotlib	2D plotting
seaborn	Statistical graphics
plotnine	ggplot2-style plots
xarray	Labeled N-dimensional arrays
zarr	Compressed array storage

Interactive Notebooks

Package	Description
JupyterLab	Interactive notebooks

---

Notes

• PyTorch is not included in this environment. PLUMED 2.10.1 uses the standalone LibTorch 2.13 (CPU) in /sw/libtorch/2.13.0-cpu for ML-CVs; no separate PyTorch installation is needed for standard PLUMED usage. For custom ML-CV training, use a dedicated environment.
• OpenMM and openff-toolkit (force field parameterization) are available in the separate openff/1 environment (module load openff/1).