Self-referencing embedded strings (SELFIES): A 100% robust molecular string representation

By Mario Krenn, Florian Hase, AkshatKumar Nigam, Pascal Friederich, Alán Aspuru-Guzik. Machine Learning: Science and Technology, 2019

957 citations recorded by the source corpus.

The discovery of novel materials and functional molecules can help to solve some of society’s most urgent challenges, ranging from efficient energy harvesting and storage to uncovering novel pharmaceutical drug candidates. Traditionally matter engineering–generally denoted as inverse design–was based massively on human intuition and high-throughput virtual screening. The last few years have seen the emergence of significant interest in computer-inspired designs based on evolutionary or deep learning methods. The major challenge here is that the standard strings molecular representation SMILES shows substantial weaknesses in that task because large fractions of strings do not correspond to valid molecules. Here, we solve this problem at a fundamental level and introduce SELFIES (SELF-referencIng Embedded Strings), a string-based representation of molecules which is 100% robust. Every SELFIES string corresponds to a valid molecule, and SELFIES can represent every molecule. SELFIES can be directly applied in arbitrary machine learning models without the adaptation of the models; each of the generated molecule candidates is valid. In our experiments, the model’s internal memory stores two orders of magnitude more diverse molecules than a similar test with SMILES. Furthermore, as all molecules are valid, it allows for explanation and interpretation of the internal working of the generative models.

Subfields: Artificial Intelligence, Computational Theory and Mathematics, Materials Chemistry

View Paper: https://www.semanticscholar.org/paper/8338a903d8078481ff8af777475f7394d00e9d57

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Marinka Zitnik
2. Jure Leskovec
3. Kexin Huang
4. Cao Xiao
5. Jimeng Sun
6. Heng Ji
7. Inkit Padhi
8. Shengchao Liu

In Computational Theory and Mathematics

1. Tianfan Fu
2. Jimeng Sun
3. Marinka Zitnik
4. Jure Leskovec
5. Shengchao Liu
6. Kexin Huang
7. Yuanqi Du
8. Cao Xiao

In Materials Chemistry

1. Tianfan Fu
2. Jimeng Sun
3. Shengchao Liu
4. Marinka Zitnik
5. Yuanqi Du
6. Cao Xiao
7. Wenhao Gao
8. Jure Leskovec

Citations in the corpus

• Leveraging large language models for predictive chemistry (2024, neutral): “For machine learning applications, there has been a lot of effort to represent a chemical with unique line encodings (for example, simplified molecular-input line-entry system (SMILES) 30 or self-referencing embedded strings (SELFIES) 31,32 ).”
• Translation between Molecules and Natural Language (2022, methodology): “One limitation of our model is using SMILES strings – recent work (Krenn et al., 2020) proposes a string representation with validity guarantees.”
• Do Large Scale Molecular Language Representations Capture Important Structural Information? (2021, methodology): “Future work will also aim at improving M O LF ORMER by employing larger models as well as larger training data, using more well-designed and/or domain-specific self-supervised tasks, and by using other string based representations like SELFIES Krenn et al. (2020).”
• ChemBERTa: Large-Scale Self-Supervised Pretraining for Molecular Property Prediction (2020, methodology): “with a more robust string representation like SELFIES [15] improve performance?”
• MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design (2022, background): “1D string representation include simplified molecular-input line-entry system (SMILES) [15] and self-referencing embedded strings (SELF-IES) [16].”
• Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development (2021, methodology): “[75] Mario Krenn, Florian Häse, AkshatKumar Nigam, Pascal Friederich, and Alan Aspuru-Guzik.”
• Large Language Models on Graphs: A Comprehensive Survey (2023, neutral): “The International Chemical Identifier (InChI) [149] To fully address this problem, SELFIES [151] is introduced which consistently yields valid molecular graphs.”