Molecular Dynamics · foundry-ml
elwood_md_v1-2
Molecular-dynamics simulation dataset mapping monomer SMILES strings to physical and thermodynamic properties.
What elwood_md_v1-2 is
Project Elwood is a molecular dynamics dataset designed for molecular property prediction in chemistry and materials science. Developed by L. Schneider, M. Schwarting, J. Mysona, H. Liang, M. Han, P. Rauscher, J. Ting, S. Venkatram, R. Ross, K. Schmidt, B. Blaiszik, I. Foster, and J. de Pablo, the collection tabulates simulated physical characteristics against molecular chemical structures.
The dataset links canonical SMILES strings to multiple molecular dynamics simulation outputs, enabling machine learning models to learn relationships between monomer structures and bulk or conformational behaviours.
How it was made
The dataset provides tabular simulation records where each entry uses a canonical Simplified Molecular Input Line Entry System (SMILES) string as its input identifier. Each chemical structure is associated with target parameters computed through molecular dynamics simulations.
The targets comprise four physical and thermodynamic measurements: cohesive energy (E_coh, measured in MPa), glass transition temperature (T_g, in Kelvin), squared radius of gyration (R_gyr, in square ångströms), and density (in kg/m³). All records are arranged under a single training split.
Key facts
- Input feature
- Canonical SMILES string
- Target properties
- Cohesive energy (MPa), glass transition temperature (K), squared radius of gyration (Ų), density (kg/m³)
- Data format
- Parquet
- Data type
- Tabular
- Splits
- train
- Publication year
- 2022
- Publisher
- Materials Data Facility
- DOI
- 10.18126/8p6m-e135
What you can use it for
- Glass transition temperature prediction
- Training regression models to predict simulated glass transition temperatures (T_g) directly from monomer chemical structures.
- Cohesive energy estimation
- Fitting machine-learning algorithms to infer cohesive energy values (E_coh) from molecular SMILES strings for monomer screening.
- Conformational and density modelling
- Evaluating models that predict monomer mass densities and squared radius of gyration (R_gyr) metrics from structural identifiers.
Explore the data
The schema, real rows, full-text search and a SQL filter — read in your browser, nothing downloaded.
Explore
Read the real rows without downloading anything
Reading rows…
LiveRead from foundry-ml/elwood_md_v1-2 at the moment you asked. Nothing is cached or stored — every row above came from that request.
How the columns are distributed
Over 410 rows of default/train
Densities (kg/m^3)
float- min
- 157.17
- median
- 1132.26
- mean
- 1129.89
- max
- 1842.11
E_coh (MPa)
float- min
- 48.36
- median
- 550.25
- mean
- 560.20
- max
- 1325.91
R_gyr (A^2)
float- min
- 0
- median
- 0
- mean
- 0
- max
- 0
SMILES
string_text- min
- 4
- median
- 26
- mean
- 29.80
- max
- 85
T_g (K)
float- min
- 196.49
- median
- 319.64
- mean
- 317.88
- max
- 754.46
Strengths and limits
Good at
- Provides four distinct simulated physical and thermodynamic endpoints for every chemical structure.
- Employs canonical SMILES strings for straightforward integration into computational chemistry and cheminformatics workflows.
- Supplied in a structured tabular format suited to regression and classification tasks.
Watch out for
- Supplies only a single training split, requiring researchers to partition validation and test sets manually.
- Contains exclusively simulated molecular dynamics outputs rather than experimentally measured physical values.
- Target variables are restricted to four specific bulk and conformational properties.
Questions
Can I look inside elwood_md_v1-2 without downloading it?
Yes. This page reads its schema, real rows and per-column statistics in the browser, with full-text search and a SQL filter. Nothing is downloaded to your machine.
What licence is elwood_md_v1-2 under?
Its authors declare cc-by-4.0. Read the terms on the original page before reusing the data.
What target variables are available in the dataset?
The dataset contains four simulated targets: cohesive energy in MPa, glass transition temperature in Kelvin, squared radius of gyration in square ångströms, and density in kg/m³.
What format is used to represent the molecules?
Molecules are represented as canonical SMILES strings in the input column.
What machine-learning tasks does the dataset support?
It is configured for tabular regression and tabular classification workflows targeting simulated monomer properties.
How are the data splits organised?
The dataset is structured with a single split labelled train.
How to cite it
Schneider, L., Schwarting, M., Mysona, J., Liang, H., Han, M., Rauscher, P., Ting, J., Venkatram, S., Ross, R., Schmidt, K., Blaiszik, B., Foster, I., & de Pablo, J. (2022). Project Elwood: MD Simulated Monomer Properties. Materials Data Facility. https://doi.org/10.18126/8p6m-e135
Written from elwood_md_v1-2's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.