File size: 6,379 Bytes
c703bc8
 
 
 
63d9b78
 
 
 
 
 
 
 
 
 
 
c703bc8
 
63d9b78
 
 
 
c703bc8
 
 
63d9b78
c703bc8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
63d9b78
 
 
 
 
 
c703bc8
 
63d9b78
 
 
 
c703bc8
 
63d9b78
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
# Model documentation & parameters

## Parameters

### Property
The supported properties are:
  - `Metal NonMetal Classifier`: Predicted by a RF model (WHICH? )
  - `Metal Semiconductor Classifier`: Classifying whether a metal could be a semiconductor. Predicted with CGCNN (ToDo: Add Ref!)
  - `Poisson Ratio`:  ToDo: Description + Reference
  - `Shear Moduli` ...
  - `Bulk Moduli`
  - `Fermi Energy`
  - `Band Gap`
  - `Absolute Energy`
  - `Formation Energy`


### Input file for crystal model
The file with information about the metal. Dependent on the property you want to predict, the format of the file differs:
- `Metal NonMetal Classifier`. It requires a single `.csv` file with the metal (chemical formula) in the first column and the crystal system in the second.
- **All others**: Predicted with CGCNN. The input can either be a single `.cif` file (to predict a single metal) or a `.zip` folder which contains multiple `.cif` (for batch prediction)



# Model card - CGCNN

**Model Details**: The [Regression Transformer](https://arxiv.org/abs/2202.01338) is a multitask Transformer that reformulates regression as a conditional sequence modeling task. This yields a dichotomous language model that seamlessly integrates property prediction with property-driven conditional generation.

**Developers**: Jannis Born and Matteo Manica from IBM Research.

**Distributors**: Original authors' code wrapped and distributed by GT4SD Team (2023) from IBM Research.

**Model date**: Preprint released in 2022, currently under review at *Nature Machine Intelligence*.

**Algorithm version**: Models trained and distributed by the original authors.
- **Molecules: QED**: Model trained on 1.6M molecules (SELFIES) from ChEMBL and their QED scores.
- **Molecules: Solubility**: QED model finetuned on the ESOL dataset from [Delaney et al (2004), *J. Chem. Inf. Comput. Sci.*](https://pubs.acs.org/doi/10.1021/ci034243x) to predict water solubility. Model trained on augmented SELFIES.
- **Molecules: USPTO**: Model trained on 2.8M [chemical reactions](https://figshare.com/articles/dataset/Chemical_reactions_from_US_patents_1976-Sep2016_/5104873) from the US patent office. The model used SELFIES and a synthetic property (total molecular weight of all precursors).
- **Molecules: Polymer**: Model finetuned on 600 ROPs (ring-opening polymerizations) with monomer-catalyst pairs. Model used three properties: conversion (`<conv>`), PDI (`<pdi>`) and Molecular Weight (`<molwt>`). Model trained with augmented SELFIES, optimized only to generate catalysts, given a monomer and the property constraints. See the example for details.
- **Molecules: Cosmo_acdl**: Model finetuned on 56k molecules with two properties (*pKa_ACDL* and *pKa_COSMO*). Model used augmented SELFIES.
- **Molecules: Pfas**: Model finetuned on ~1k PFAS (Perfluoroalkyl and Polyfluoroalkyl Substances) molecules with 9 properties including some experimentally measured ones (biodegradability, LD50 etc) and some synthetic ones (SCScore, molecular weight). Model trained on augmented SELFIES.
- **Molecules: Logp_and_synthesizability**: Model trained on 2.9M molecules (SELFIES) from PubChem with **two** synthetic properties, the logP (partition coefficient) and the [SCScore by Coley et al. (2018); *J. Chem. Inf. Model.*](https://pubs.acs.org/doi/full/10.1021/acs.jcim.7b00622?casa_token=JZzOrdWlQ_QAAAAA%3A3_ynCfBJRJN7wmP2gyAR0EWXY-pNW_l-SGwSSU2SGfl5v5SxcvqhoaPNDhxq4THberPoyyYqTZELD4Ck)
- **Molecules: Crippen_logp**: Model trained on 2.9M molecules (SMILES) from PubChem, but *only* on logP (partition coefficient).
- **Proteins: Stability**: Model pretrained on 2.6M peptides from UniProt with the Boman index as property. Finetuned on the [**Stability**](https://www.science.org/doi/full/10.1126/science.aan0693) dataset from the [TAPE benchmark](https://proceedings.neurips.cc/paper/2019/hash/37f65c068b7723cd7809ee2d31d7861c-Abstract.html) which has ~65k samples.

**Model type**: A Transformer-based language model that is trained on alphanumeric sequence to simultaneously perform sequence regression or conditional sequence generation.

**Information about training algorithms, parameters, fairness constraints or other applied approaches, and features**: 
All models are trained with an alternated training scheme that alternated between optimizing the cross-entropy loss on the property tokens ("regression") or the self-consistency objective on the molecular tokens. See the [Regression Transformer](https://arxiv.org/abs/2202.01338) paper for details.

**Paper or other resource for more information**: 
The [Regression Transformer](https://arxiv.org/abs/2202.01338) paper. See the [source code](https://github.com/IBM/regression-transformer) for details.

**License**: MIT

**Where to send questions or comments about the model**: Open an issue on [GT4SD repository](https://github.com/GT4SD/gt4sd-core).

**Intended Use. Use cases that were envisioned during development**: Chemical research, in particular drug discovery.

**Primary intended uses/users**: Researchers and computational chemists using the model for model comparison or research exploration purposes.

**Out-of-scope use cases**: Production-level inference, producing molecules with harmful properties.

**Factors**: Not applicable.

**Metrics**: High predictive power for the properties of that specific algorithm version.

**Datasets**: Different ones, as described under **Algorithm version**.

**Ethical Considerations**: No specific considerations as no private/personal data is involved. Please consult with the authors in case of questions.

**Caveats and Recommendations**: Please consult with original authors in case of questions.

Model card prototype inspired by [Mitchell et al. (2019)](https://dl.acm.org/doi/abs/10.1145/3287560.3287596?casa_token=XD4eHiE2cRUAAAAA:NL11gMa1hGPOUKTAbtXnbVQBDBbjxwcjGECF_i-WC_3g1aBgU1Hbz_f2b4kI_m1in-w__1ztGeHnwHs)


# Model card - RandomForestMetalClassifier

ToDo...


# Citation

```bib
@article{manica2022gt4sd,
  title={GT4SD: Generative Toolkit for Scientific Discovery},
  author={Manica, Matteo and Cadow, Joris and Christofidellis, Dimitrios and Dave, Ashish and Born, Jannis and Clarke, Dean and Teukam, Yves Gaetan Nana and Hoffman, Samuel C and Buchan, Matthew and Chenthamarakshan, Vijil and others},
  journal={arXiv preprint arXiv:2207.03928},
  year={2022}
}
```