Computer-Assisted Retrosynthesis Based on Molecular Similarity

By Connor W. Coley, Luke Rogers, William H. Green Jr., Klavs F. Jensen. 2017

381 citations recorded by the source corpus.

We demonstrate molecular similarity to be a surprisingly effective metric for proposing and ranking one-step retrosynthetic disconnections based on analogy to precedent reactions. The developed approach mimics the retrosynthetic strategy defined implicitly by a corpus of known reactions without the need to encode any chemical knowledge. Using 40 000 reactions from the patent literature as a knowledge base, the recorded reactants are among the top 10 proposed precursors in 74.1% of 5000 test reactions, providing strong quantitative support for our methodology. Extension of the one-step strategy to multistep pathway planning is demonstrated and discussed for two exemplary drug products.

Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology, Computational Chemistry & Drug Discovery

How researchers see this paper

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

In Computational Theory and Mathematics

In Materials Chemistry

In Molecular Biology

In Computational Chemistry & Drug Discovery

Citations in the corpus

Modern Hopfield Networks for Few- and Zero-Shot Reaction Prediction (2021, methodology): “More recently, machine learning and deep learning models have been used to learn the rules of chemical reactions and, thereby, enhancing CASP systems (Coley et al., 2018; Segler et al., 2018; Coley et al., 2017).”
Emerging materials intelligence ecosystems propelled by machine learning (2020, methodology): “To this end, a tool called retrosim 213 has been developed that uses similarities in the fingerprints of molecules and a library of reactions to select the rules to apply.”
Artificial intelligence in drug discovery: recent advances and future perspectives (2021, methodology): “This separation, inspired by a previously reported molecular similarity method [172], was shown to achieve performance gains over previous baselines.”
State-of-the-art augmented NLP transformer models for direct and single-step retrosynthesis (2020, methodology): “Meanwhile other approaches based on similarity [20], convolutional [21, 22, 23], and graphs show promising results [24, 25].”
Mapping the space of chemical reactions using attention-based neural networks (2020, methodology): “The algorithm consists of four steps: (1) LSH Forest-based indexing, (2) k-nearest-neighbour graph generation, (3) minimum spanning tree calculation using Kurskal’s algorithm and (4) creating the tree-like layout.”