Repository navigation
Overview
This is a minor release of MDAnalysis.
This release supports NumPy 2.0+ and offers backwards compatibility through to NumPy 1.26.0.
Supported Python versions:
- 3.11, 3.12, 3.13, 3.14
Major features & changes
See the CHANGELOG for more details.
Analysis methods
- Paralllelization supporrt for; DiffusionMap (PR #4745), AtomicDistances (PR #4822),
- DSSP improvements
- Fixed issue where Analysis classes would not work with GSD trajectories (Issue #5224, PR #5260).
- AtomicDistances now uses the
resultsdictionary for its outputs, in line with other analysis classes (PR #5347). - Fix to mean aggregation of InterRDF bins when running in parallel (PR #5344).
- Fix to issue where numpy boolean arrays passed to
run()would incorrectly select frame indices (PPR #5476). - Fix to incorrect reduction of eigenvectors in the
GNMAnalysisclass when running analysis with multiprocessing (PR #5495). - Fix incorrect aggreggation in
InterRDF_swhen running in parallel (PR #5497).
File formats
- TPR reader improvements
- Support for GROMACS 2026 TPR files (PR #5219).
- Fix issue with reading old TPR files (TPX version < 103) (PR #5146).
- Fix to position & velocity parsing to properly convert to MDAnalysis base units rather than using Gromacs' native units (PR #5374).
- Support for reading linear angle potentials (PR #5372).
- Box vectors can now be read from TPR files (PR #5466).
- Implementation of a
mda.fetch_pdb()convenience which allows for creation ofUniversesby directly pulling from the Protein DataBank (PR #4943). - Added support for reading unit cell dimensions from CRYSIN records in MOL2 files (PR #5187).
- Implementation of an MMCIF file reader using the
gemmilibrary (PR #4712).
Library internals
- Significant
AtomGroup.wrap()performance improvement.wrap()now scales close to O(N) instead of O(N^2) at large compound counts. - Bond guessing now words with mixed case atomtypes (PR #5343).
- Velocities and forces are now also rotated when applying trajectory rotation transformations (PR #5421).
- Fix inccorrect
capped_distancein some triclinic boxes (where bothb_xandc_yare non-zero), leading to atom pairs within the cutoff distance being missed (PR #5481).
New Contributors
- @spyke7 made their first contribution in #5153
- @Pranay096 made their first contribution in #5158
- @PardhavMaradani made their first contribution in #5185
- @tani-dubey made their first contribution in #5173
- @BradyAJohnston made their first contribution in #5202
- @IAyaanHere made their first contribution in #5205
- @khushi-phougat made their first contribution in #5212
- @Dreamstick9 made their first contribution in #5156
- @jeremyleung521 made their first contribution in #5237
- @harshitgajjela-droid made their first contribution in #5265
- @kunjsinha made their first contribution in #5260
- @aygarwal made their first contribution in #5277
- @ParthUppal523 made their first contribution in #5294
- @ollyfutur made their first contribution in #5296
- @Amarendra22 made their first contribution in #5291
- @shubhammittal2241 made their first contribution in #5178
- @charity-g made their first contribution in #5347
- @saiudayagiri made their first contribution in #5401
- @apoorva-01 made their first contribution in #5415
- @Aryaman-Chaudhri made their first contribution in #5139
- @sici17 made their first contribution in #5439
- @AkshitBoora made their first contribution in #5466
Full Changelog: package-2.10.0...package-2.11.0