CoolFace
Datasetpublic

niklastr/microsolvated_peptides

Microsolvated Peptide Ensembles This repository provides explicit-water microsolvated structures for a subset of peptides initialized from ManyPeptidesMD conformers. Overview Path Model Use trajectories/v1/ Amber14 in TIP3P, cropped to the nearest 128 waters Peptide-solvent sampling webdatasets/v1/train_300K/ Streamable samples from the 300 K state Model training webdatasets/v1/train_300-500K/ Streamable samples from all REMD states… See the full description on the dataset page: https://huggingface.co/datasets/niklastr/microsolvated_peptides.

sourceHugging Facecc-by-4.0updated 2mo agoView on Hugging Face
4likes5.3kdownloads
Dataset Card

Microsolvated Peptide Ensembles

This repository provides explicit-water microsolvated structures for a subset of peptides initialized from ManyPeptidesMD conformers.

Overview

PathModelUse
trajectories/v1/Amber14 in TIP3P, cropped to the nearest 128 watersPeptide-solvent sampling
webdatasets/v1/train_300K/Streamable samples from the 300 K stateModel training
webdatasets/v1/train_300-500K/Streamable samples from all REMD statesTemperature-conditioned training
webdatasets/v1/train_300K_neutral/Neutral-peptide subset at 300 KModel training
webdatasets/v1/train_300-500K_neutral/Neutral-peptide subset from all REMD statesTemperature-conditioned training

Files follow:

text
trajectories/v1/<split>/<length>/<sequence>/state_YY_<temperature>K.npz

Each NPZ contains positions, smiles, id, temperature, potential energy, and metadata. positions has shape (frames, atoms, 4) with rows [atomic_number, x_nm, y_nm, z_nm]. Files also contain the parent periodic box_vectors.

Files aggregate independent REMD bursts initialized from up to 100 source conformers. Saved frames within each burst are 10 ps apart. Please note that the data were simulated using REMD, which breaks temporal ordering.

Simulation

Starting structures came from the corresponding ManyPeptidesMD trajectories. Up to 100 conformers per sequence were selected by uniform stride from the end of each source trajectory. Peptide sequences, atom order, and train/validation/test assignments follow ManyPeptidesMD.

Replica Exchange

ParameterValue
Replicas8, geometrically spaced from 300 to 500 K
Temperatures300.00, 322.71, 347.14, 373.42, 401.69, 432.10, 464.81, 500.00 K
IntegratorLangevin, 2 fs, 0.3 ps^-1 collision rate
Warm-up / production0.5 ns / 100 ns per replica
Save interval5,000 steps = 10 ps
Exchange interval5,000 steps = 10 ps

Outputs were demultiplexed by thermodynamic state. Each state file therefore represents a fixed temperature, not one replica's path through the ladder. Select state_00_300p00K.npz for 300 K data.

Energy Function

  • Peptide: Amber14.
  • Explicit Solvent: TIP3P; PME at 0.9 nm; H-bond constraints. Rectangular box with 1.0 nm solvent padding.
  • Microsolvated energy Scoring: finite Amber14/TIP3P cluster; reaction field at 2.0 nm with dielectric 78.5.

Microsolvated Frames

Explicit-water MD ran in full periodic TIP3P boxes. No NPT adjustment of the box dimensions was made. The distributed trajectory frames are compact local environments generated as follows:

  1. 1.Reimage the peptide and counterions in the parent periodic box.
  2. 2.Rank waters independently in each frame by center-of-mass distance to the nearest peptide-or-ion atom.
  3. 3.Retain the nearest 128 complete TIP3P waters.
  4. 4.Store atoms as peptide[, counterions] + 128 x [O, H, H] together with the parent box vectors.

The cropped cluster omits the bulk environment. Do not evaluate it with the original periodic PME energy or use it for bulk-water observables. Stored energy values are recomputed for the finite microsolvated cluster with the reaction-field model above.

Citation

Please cite this dataset and ManyPeptidesMD when using its peptide sequences or starting conformers:

bibtex
@misc{rindtorff_2026_microsolvated_peptide_trajectories,
  author       = {Rindtorff, Niklas},
  title        = {Microsolvated Peptide Ensembles},
  year         = {2026},
  publisher    = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/niklastr/microsolvated_peptides}},
  note         = {Dataset in progress; accessed 2026-07-23}
}

Source dataset: ManyPeptidesMD.

License

This derived dataset is released under CC BY 4.0. ManyPeptidesMD is distributed under the terms stated in its dataset repository. See the source repository for authoritative terms and provenance caveats.

niklastr/microsolvated_peptides · CoolFace