Commit
·
2d68ddf
1
Parent(s):
66af513
updates the Readme file
Browse files
README.md
CHANGED
|
@@ -1,55 +1,56 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
| 3 |
-
- cc-by-4.0
|
| 4 |
license_link: https://creativecommons.org/licenses/by/4.0
|
| 5 |
tags:
|
| 6 |
-
- homo-lumo-gaps
|
| 7 |
-
- energy
|
| 8 |
-
- dipole-moments
|
| 9 |
-
- quantum-chemistry
|
| 10 |
-
- pubchem
|
| 11 |
-
- small-molecules
|
| 12 |
annotations_creators:
|
| 13 |
-
- crowdsourced
|
| 14 |
pretty_name: pubchemqc-b3lyp
|
| 15 |
size_categories:
|
| 16 |
-
- 10K<n<100M
|
| 17 |
source_datasets:
|
| 18 |
-
- pubchem
|
| 19 |
-
- pubchemqc-b3lyp
|
| 20 |
task_categories:
|
| 21 |
-
- tabular-regression
|
| 22 |
-
- other
|
| 23 |
task_ids:
|
| 24 |
-
- tabular-single-column-regression
|
| 25 |
configs:
|
| 26 |
-
- config_name: b3lyp_pm6
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
- config_name: b3lyp_pm6_chon300nosalt
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
- config_name: b3lyp_pm6_chon500nosalt
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
- config_name: b3lyp_pm6_chnopsfcl300nosalt
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
- config_name: b3lyp_pm6_chnopsfcl500nosalt
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
- config_name: b3lyp_pm6_chnopsfclnakmgca500
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
---
|
| 52 |
-
|
|
|
|
| 53 |
|
| 54 |
## Table of Contents
|
| 55 |
|
|
@@ -61,6 +62,9 @@ configs:
|
|
| 61 |
- [Data Instances](#data-instances)
|
| 62 |
- [Data Fields](#data-fields)
|
| 63 |
- [Data Splits and Configurations](#data-splits-and-configurations)
|
|
|
|
|
|
|
|
|
|
| 64 |
- [Dataset Creation](#dataset-creation)
|
| 65 |
- [Curation Rationale](#curation-rationale)
|
| 66 |
- [Source Data](#source-data)
|
|
@@ -85,12 +89,12 @@ configs:
|
|
| 85 |
|
| 86 |
### Dataset Summary
|
| 87 |
|
| 88 |
-
The presented **PubChemQC B3LYP/6-31G
|
| 89 |
properties of 85,938,443 molecules, encompassing a broad spectrum of molecules from
|
| 90 |
essential compounds to biomolecules with a molecular weight up to 1000. These molecules
|
| 91 |
account for 94.0% of the original PubChem Compound catalog as of August 29, 2016.
|
| 92 |
The electronic properties, including orbitals, orbital energies, total energies, dipole
|
| 93 |
-
moments, and other pertinent properties, were computed by using the B3LYP/6-31G
|
| 94 |
methods.
|
| 95 |
|
| 96 |
## Dataset Structure
|
|
@@ -194,87 +198,145 @@ An example of a data instance is as follows:
|
|
| 194 |
|
| 195 |
### Data Fields
|
| 196 |
|
| 197 |
-
|
| 198 |
-
|
| 199 |
-
|
| 200 |
-
|
| 201 |
-
|
| 202 |
-
|
| 203 |
-
|
| 204 |
-
|
| 205 |
-
|
| 206 |
-
|
| 207 |
-
|
| 208 |
-
|
| 209 |
-
|
| 210 |
-
|
| 211 |
-
|
| 212 |
-
|
| 213 |
-
|
| 214 |
-
|
| 215 |
-
|
| 216 |
-
|
| 217 |
-
|
| 218 |
-
|
| 219 |
-
|
| 220 |
-
|
| 221 |
-
|
| 222 |
-
|
| 223 |
-
|
| 224 |
-
|
| 225 |
-
|
| 226 |
-
|
| 227 |
-
|
| 228 |
-
|
| 229 |
-
|
| 230 |
-
|
| 231 |
-
|
| 232 |
-
|
| 233 |
-
|
| 234 |
-
|
| 235 |
-
|
| 236 |
-
|
| 237 |
-
|
| 238 |
-
|
| 239 |
|
| 240 |
### Data Splits and Configurations
|
| 241 |
|
| 242 |
-
The dataset has only one
|
| 243 |
has six configurations/subsets:
|
| 244 |
-
|
| 245 |
-
-
|
| 246 |
-
-
|
| 247 |
-
-
|
| 248 |
-
-
|
| 249 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 250 |
|
| 251 |
## Dataset Creation
|
| 252 |
|
| 253 |
### Curation Rationale
|
| 254 |
|
| 255 |
-
The present version of PubChemQC B3LYP/6-31G
|
| 256 |
from its original Postgresql database, transformed into a dictionary and
|
| 257 |
-
stored in the
|
| 258 |
|
| 259 |
### Source Data
|
| 260 |
|
| 261 |
-
The link to the original PubChemQC B3LYP/6-31G
|
| 262 |
found [here](https://nakatamaho.riken.jp/pubchemqc.riken.jp/b3lyp_pm6_datasets.html)
|
| 263 |
|
| 264 |
#### Initial Data Collection and Normalization
|
| 265 |
|
| 266 |
Other than the changes detailed in Sec. [Curation Rationale](#curation-rationale),
|
| 267 |
-
no data modification has been performed on the PubChemQC B3LYP/6-31G
|
| 268 |
|
| 269 |
### Personal and Sensitive Information
|
| 270 |
|
| 271 |
-
The PubChemQC B3LYP/6-31G
|
| 272 |
|
| 273 |
## Considerations for Using the Data
|
| 274 |
|
| 275 |
### Social Impact of Dataset
|
| 276 |
|
| 277 |
-
The PubChemQC B3LYP/6-31G
|
| 278 |
|
| 279 |
## Additional Information
|
| 280 |
|
|
@@ -298,7 +360,7 @@ The PubChemQC B3LYP/6-31G*//PM6 dataset paves the way for applications in drug d
|
|
| 298 |
journal = {Journal of Chemical Information and Modeling},
|
| 299 |
pages = {5734-5754},
|
| 300 |
publisher = {American Chemical Society},
|
| 301 |
-
title = {{PubChemQC B3LYP/6-31G*//PM6 Data Set: The Electronic Structures of 86 Million Molecules
|
| 302 |
Using B3LYP/6-31G* Calculations}},
|
| 303 |
volume = {63},
|
| 304 |
url = {https://pubs.acs.org/doi/abs/10.1021/acs.jcim.3c00899},
|
|
@@ -308,4 +370,4 @@ The PubChemQC B3LYP/6-31G*//PM6 dataset paves the way for applications in drug d
|
|
| 308 |
|
| 309 |
### Contributions
|
| 310 |
|
| 311 |
-
- **Mohammad Mostafanejad**, The Molecular Sciences Software Institute (MolSSI)
|
|
|
|
| 1 |
---
|
| 2 |
+
license:
|
| 3 |
+
- cc-by-4.0
|
| 4 |
license_link: https://creativecommons.org/licenses/by/4.0
|
| 5 |
tags:
|
| 6 |
+
- homo-lumo-gaps
|
| 7 |
+
- energy
|
| 8 |
+
- dipole-moments
|
| 9 |
+
- quantum-chemistry
|
| 10 |
+
- pubchem
|
| 11 |
+
- small-molecules
|
| 12 |
annotations_creators:
|
| 13 |
+
- crowdsourced
|
| 14 |
pretty_name: pubchemqc-b3lyp
|
| 15 |
size_categories:
|
| 16 |
+
- 10K<n<100M
|
| 17 |
source_datasets:
|
| 18 |
+
- pubchem
|
| 19 |
+
- pubchemqc-b3lyp
|
| 20 |
task_categories:
|
| 21 |
+
- tabular-regression
|
| 22 |
+
- other
|
| 23 |
task_ids:
|
| 24 |
+
- tabular-single-column-regression
|
| 25 |
configs:
|
| 26 |
+
- config_name: b3lyp_pm6
|
| 27 |
+
data_files:
|
| 28 |
+
- split: train
|
| 29 |
+
path: "data/b3lyp_pm6/train/*.json"
|
| 30 |
+
default: true
|
| 31 |
+
- config_name: b3lyp_pm6_chon300nosalt
|
| 32 |
+
data_files:
|
| 33 |
+
- split: train
|
| 34 |
+
path: "data/b3lyp_pm6_chon300nosalt/train/*.json"
|
| 35 |
+
- config_name: b3lyp_pm6_chon500nosalt
|
| 36 |
+
data_files:
|
| 37 |
+
- split: train
|
| 38 |
+
path: "data/b3lyp_pm6_chon500nosalt/train/*.json"
|
| 39 |
+
- config_name: b3lyp_pm6_chnopsfcl300nosalt
|
| 40 |
+
data_files:
|
| 41 |
+
- split: train
|
| 42 |
+
path: "data/b3lyp_pm6_chnopsfcl300nosalt/train/*.json"
|
| 43 |
+
- config_name: b3lyp_pm6_chnopsfcl500nosalt
|
| 44 |
+
data_files:
|
| 45 |
+
- split: train
|
| 46 |
+
path: "data/b3lyp_pm6_chnopsfcl500nosalt/train/*.json"
|
| 47 |
+
- config_name: b3lyp_pm6_chnopsfclnakmgca500
|
| 48 |
+
data_files:
|
| 49 |
+
- split: train
|
| 50 |
+
path: "data/b3lyp_pm6_chnopsfclnakmgca500/train/*.json"
|
| 51 |
---
|
| 52 |
+
|
| 53 |
+
# PubChemQC-B3LYP/6-31G\*//PM6 Dataset
|
| 54 |
|
| 55 |
## Table of Contents
|
| 56 |
|
|
|
|
| 62 |
- [Data Instances](#data-instances)
|
| 63 |
- [Data Fields](#data-fields)
|
| 64 |
- [Data Splits and Configurations](#data-splits-and-configurations)
|
| 65 |
+
- [How to Use the Dataset](#how-to-use-the-dataset)
|
| 66 |
+
- [Prerequisites](#prerequisites)
|
| 67 |
+
- [Accessing the Data](#accessing-the-data)
|
| 68 |
- [Dataset Creation](#dataset-creation)
|
| 69 |
- [Curation Rationale](#curation-rationale)
|
| 70 |
- [Source Data](#source-data)
|
|
|
|
| 89 |
|
| 90 |
### Dataset Summary
|
| 91 |
|
| 92 |
+
The presented **PubChemQC B3LYP/6-31G\*//PM6** data set is composed of the electronic
|
| 93 |
properties of 85,938,443 molecules, encompassing a broad spectrum of molecules from
|
| 94 |
essential compounds to biomolecules with a molecular weight up to 1000. These molecules
|
| 95 |
account for 94.0% of the original PubChem Compound catalog as of August 29, 2016.
|
| 96 |
The electronic properties, including orbitals, orbital energies, total energies, dipole
|
| 97 |
+
moments, and other pertinent properties, were computed by using the B3LYP/6-31G\* and PM6
|
| 98 |
methods.
|
| 99 |
|
| 100 |
## Dataset Structure
|
|
|
|
| 198 |
|
| 199 |
### Data Fields
|
| 200 |
|
| 201 |
+
| Field | Description |
|
| 202 |
+
| ------------------------------- | ------------------------------------------------------------------------------------------------------------------------------ |
|
| 203 |
+
| cid | Pubchem Compound ID |
|
| 204 |
+
| state | Electronic state |
|
| 205 |
+
| pubchem-inchi | InChI extracted from PubChem Compound entry |
|
| 206 |
+
| pubchem-charge | Molecular charge extracted from PubChem Compound entry |
|
| 207 |
+
| pubchem-version | PubChem Compound database version |
|
| 208 |
+
| name | Name of the input file used for the B3LYP/6-31G\*//PM6 calculation |
|
| 209 |
+
| coordinates | Cartesian coordinates of the molecular geometry optimized with PM6 method in Angstroem |
|
| 210 |
+
| atomic-numbers | An array of atomic numbers |
|
| 211 |
+
| atom-count | Number of atoms in the molecule |
|
| 212 |
+
| heavy-atom-count | Number of heavy atoms in the molecule |
|
| 213 |
+
| core-electrons | Number of core electrons in each atom's pseudopotentials |
|
| 214 |
+
| bond-order | Bond order |
|
| 215 |
+
| connection-indices | Connection indices between atoms |
|
| 216 |
+
| formula | Chemical formula |
|
| 217 |
+
| version | Version number |
|
| 218 |
+
| obabel-inchi | InChI of the structure generated by Open Babel |
|
| 219 |
+
| pm6-obabel-canonical-smiles | Canonical SMILES for the structure generated by Open Babel |
|
| 220 |
+
| charge | Molecular charge |
|
| 221 |
+
| energy-beta-gap | HOMO-LUMO energy gap for beta spin orbitals |
|
| 222 |
+
| energy-beta-homo | Energy of the highest-occupied molecular orbital (HOMO) of beta spin symmetry |
|
| 223 |
+
| energy-beta-lumo | Energy of the lowest-unoccupied molecular orbital (LUMO) of beta spin symmetry |
|
| 224 |
+
| energy-alpha-gap | HOMO-LUMO energy gap for alpha spin orbitals |
|
| 225 |
+
| energy-alpha-homo | Energy of the highest-occupied molecular orbital (HOMO) of alpha spin symmetry |
|
| 226 |
+
| energy-alpha-lumo | Energy of the lowest-unoccupied molecular orbital (LUMO) of alpha spin symmetry |
|
| 227 |
+
| total-energy | Total electronic energy of the molecule calculated at the B3LYP/6-31G\* level of theory |
|
| 228 |
+
| homos | 1D index array of the highest occupied molecular orbital (HOMO) with one (two) element(s) for the (un)restricted wavefunctions |
|
| 229 |
+
| orbital-energies | 1D array of orbital energies in hartree with one (two) member(s) for the (un)restricted wavefunction |
|
| 230 |
+
| mo-count | Number of molecular orbitals |
|
| 231 |
+
| basis-count | Number of basis functions |
|
| 232 |
+
| multiplicity | Spin multiplicity |
|
| 233 |
+
| molecular-mass | Molecular mass |
|
| 234 |
+
| number-of-atoms | Number of atoms in the molecule |
|
| 235 |
+
| lowdin-partial-charges | Lowdin partial atomic charges |
|
| 236 |
+
| mulliken-partial-charges | Mulliken partial atomic charges |
|
| 237 |
+
| dipole-moment | Dipole moment |
|
| 238 |
+
| pubchem-multiplicity | Spin multiplicity of the molecule extracted from PubChem Compound |
|
| 239 |
+
| pubchem-obabel-canonical-smiles | Canonical SMILES of the molecule extracted from PubChem Compound generated by Open Babel |
|
| 240 |
+
| pubchem-isomeric-smiles | Isomeric SMILES of the molecule extracted from PubChem Compound calculated by the OpenEye's OEChem Toolkit |
|
| 241 |
+
| pubchem-molecular-weight | Molecular weight extracted from the PubChem Compound entry |
|
| 242 |
+
| pubchem-molecular-formula | Molecular formula extracted from the PubChem Compound entry |
|
| 243 |
|
| 244 |
### Data Splits and Configurations
|
| 245 |
|
| 246 |
+
The dataset has only one `train` split. The PubChemQC B3LYP/6-31G\*//PM6 dataset
|
| 247 |
has six configurations/subsets:
|
| 248 |
+
|
| 249 |
+
- `b3lyp_pm6` (default)
|
| 250 |
+
- `b3lyp_pm6_chon300nosalt`
|
| 251 |
+
- `b3lyp_pm6_chon500nosalt`
|
| 252 |
+
- `b3lyp_pm6_chnopsfcl300nosalt`
|
| 253 |
+
- `b3lyp_pm6_chnopsfcl500nosalt`
|
| 254 |
+
- `b3lyp_pm6_chnopsfclnakmgca500`
|
| 255 |
+
|
| 256 |
+
## How to Use the Dataset
|
| 257 |
+
|
| 258 |
+
### Prerequisites
|
| 259 |
+
|
| 260 |
+
We recommend isolating your work in a virtualenv or conda environment.
|
| 261 |
+
You can create a new conda environment, `pubchemqc`,
|
| 262 |
+
|
| 263 |
+
```bash
|
| 264 |
+
conda create -n pubchemqc python=3.12
|
| 265 |
+
```
|
| 266 |
+
|
| 267 |
+
and activate it using the following command
|
| 268 |
+
|
| 269 |
+
```bash
|
| 270 |
+
conda activate pubchemqc
|
| 271 |
+
```
|
| 272 |
+
|
| 273 |
+
Once the conda environment is activated, you can
|
| 274 |
+
install the dependencies in it as shown below
|
| 275 |
+
|
| 276 |
+
```bash
|
| 277 |
+
pip install huggingface_hub ijson
|
| 278 |
+
```
|
| 279 |
+
|
| 280 |
+
### Accessing the Data
|
| 281 |
+
|
| 282 |
+
Once the required packages are installed, you can run the following code
|
| 283 |
+
to access the data
|
| 284 |
+
|
| 285 |
+
```python
|
| 286 |
+
# import the modules
|
| 287 |
+
from datasets import load_dataset
|
| 288 |
+
|
| 289 |
+
# load the dataset with streaming
|
| 290 |
+
hub_ds = load_dataset(path="molssiai-hub/pubchemqc-b3lyp",
|
| 291 |
+
name="b3lyp_pm6",
|
| 292 |
+
split="train",
|
| 293 |
+
streaming=True,
|
| 294 |
+
cache_dir="./tmp",
|
| 295 |
+
trust_remote_code=True)
|
| 296 |
+
|
| 297 |
+
# fetch a batch of 32 samples from the dataset
|
| 298 |
+
ds = list(hub_ds.take(32))
|
| 299 |
+
```
|
| 300 |
+
|
| 301 |
+
The argument `name` by default is set to `b3lyp_pm6` which refers to
|
| 302 |
+
the entire dataset. Other configurations (subsets), listed in
|
| 303 |
+
Sec. [Data Splits and Configurations](#data-splits-and-configurations),
|
| 304 |
+
can also be selected.
|
| 305 |
+
|
| 306 |
+
The `split` must be set to `train` as it is the only split in our dataset.
|
| 307 |
+
We recommend using `streaming=True` to avoid downloading the entire dataset
|
| 308 |
+
on disk. The `cache_dir` allows us to store the Hugging Face datasets' and
|
| 309 |
+
models' artifacts in a non-default directory (by default, it is set to
|
| 310 |
+
`~/.cache/huggingface`). As we are using a custom
|
| 311 |
+
[load script](https://huggingface.co/datasets/molssiai-hub/pubchemqc-b3lyp/blob/main/pubchemqc-b3lyp.py), the `trust_remote_code` argument should also be set to `True`.
|
| 312 |
|
| 313 |
## Dataset Creation
|
| 314 |
|
| 315 |
### Curation Rationale
|
| 316 |
|
| 317 |
+
The present version of PubChemQC B3LYP/6-31G\*//PM6 dataset has been extracted
|
| 318 |
from its original Postgresql database, transformed into a dictionary and
|
| 319 |
+
stored in the `.json` format.
|
| 320 |
|
| 321 |
### Source Data
|
| 322 |
|
| 323 |
+
The link to the original PubChemQC B3LYP/6-31G\*//PM6 dataset repository can be
|
| 324 |
found [here](https://nakatamaho.riken.jp/pubchemqc.riken.jp/b3lyp_pm6_datasets.html)
|
| 325 |
|
| 326 |
#### Initial Data Collection and Normalization
|
| 327 |
|
| 328 |
Other than the changes detailed in Sec. [Curation Rationale](#curation-rationale),
|
| 329 |
+
no data modification has been performed on the PubChemQC B3LYP/6-31G\*//PM6 dataset.
|
| 330 |
|
| 331 |
### Personal and Sensitive Information
|
| 332 |
|
| 333 |
+
The PubChemQC B3LYP/6-31G\*//PM6 dataset does not involve any personal or sensitive information.
|
| 334 |
|
| 335 |
## Considerations for Using the Data
|
| 336 |
|
| 337 |
### Social Impact of Dataset
|
| 338 |
|
| 339 |
+
The PubChemQC B3LYP/6-31G\*//PM6 dataset paves the way for applications in drug discovery and materials science, among others.
|
| 340 |
|
| 341 |
## Additional Information
|
| 342 |
|
|
|
|
| 360 |
journal = {Journal of Chemical Information and Modeling},
|
| 361 |
pages = {5734-5754},
|
| 362 |
publisher = {American Chemical Society},
|
| 363 |
+
title = {{PubChemQC B3LYP/6-31G*//PM6 Data Set: The Electronic Structures of 86 Million Molecules
|
| 364 |
Using B3LYP/6-31G* Calculations}},
|
| 365 |
volume = {63},
|
| 366 |
url = {https://pubs.acs.org/doi/abs/10.1021/acs.jcim.3c00899},
|
|
|
|
| 370 |
|
| 371 |
### Contributions
|
| 372 |
|
| 373 |
+
- **Mohammad Mostafanejad**, The Molecular Sciences Software Institute (MolSSI)
|