RetroPrime: A Diverse, plausible and Transformer-based method for Single-Step retrosynthesis predictions

Wang, Xiaorui; Li, Yuquan; Qiu, Jiezhong; Chen, Guangyong; Liu, Huanxiang; Liao, Benben; Hsieh, Chang‐Yu; Yao, Xiaojun

doi:10.1016/j.cej.2021.129845

Cited by 70 publications

(118 citation statements)

References 32 publications

Supporting

Mentioning

106

Contrasting

Order By: Relevance

“…by Lin, Neuralsym a , 2020 6 , 33 , 73 47.8 Dai et al, Graph Logic Network a , 2019 73 39.3 Liu et al,–rep. by Lin, LSTM-based, 2020 16 , 33 46.9 Genheden et al, AiZynthfinder, ANN + MCTS a , 2020 14 , 49 43–72 Transformer-based Zheng et al, SCROP, 2020 34 41.5 Wang et al, RetroPrime, 2021 48 44.1 Tetko et al, Augmented Transformer, 2020 35 46.2 Lin et al, AutoSynRoute, Transformer + MCTS, 2020 33 54.1 RetroTRAE 58.3 RetroTRAE (with SM and DM) 61.6 The results are based on either filtered MIT-full 46 , 47 or MIT-fully atom mapped 15 reaction datasets. a Reaction templates were used.…”

Section: Resultsmentioning

confidence: 99%

“…Performance differences in the SMILES-based Transformer models are attributed to improvements in data augmentation (with non-canonical SMILES) 35 , 48 , tokenization scheme (character or atom level) 31 , 33 , and postprocessing (by rectifying invalid SMILES) 32 , 34 . The better prediction accuracy of our model appears to be due to better reaction representation beyond the standard SMILES.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Retrosynthetic reaction pathway prediction through neural machine translation of atomic environments

Ucak

Ashyrmamatov

Ko³

et al. 2022

Nat Commun

View full text Add to dashboard Cite

Designing efficient synthetic routes for a target molecule remains a major challenge in organic synthesis. Atom environments are ideal, stand-alone, chemically meaningful building blocks providing a high-resolution molecular representation. Our approach mimics chemical reasoning, and predicts reactant candidates by learning the changes of atom environments associated with the chemical reaction. Through careful inspection of reactant candidates, we demonstrate atom environments as promising descriptors for studying reaction route prediction and discovery. Here, we present a new single-step retrosynthesis prediction method, viz. RetroTRAE, being free from all SMILES-based translation issues, yields a top-1 accuracy of 58.3% on the USPTO test dataset, and top-1 accuracy reaches to 61.6% with the inclusion of highly similar analogs, outperforming other state-of-the-art neural machine translation-based methods. Our methodology introduces a novel scheme for fragmental and topological descriptors to be used as natural inputs for retrosynthetic prediction tasks.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Retrosynthetic reaction pathway prediction through neural machine translation of atomic environments

Ucak

Ashyrmamatov

Ko³

et al. 2022

Nat Commun

View full text Add to dashboard Cite

show abstract

“…In this work, we strictly categorize generative models which require additional help from RDKit [1] for molecule editing into this method group. Most existing works [34,26,27,33] approach the task by a two-stage procedure. Despite their architecture differences between GNN and Transformer, they follow the same idea: They first convert the product into synthons by breaking the reactive bond via RDKit, then complete the synthons into reactants by either leaving groups selection [27], graph generation [26], or SMILES generation [34,33].…”

Section: Retrosynthesis Predictionmentioning

confidence: 99%

“…Most existing works [34,26,27,33] approach the task by a two-stage procedure. Despite their architecture differences between GNN and Transformer, they follow the same idea: They first convert the product into synthons by breaking the reactive bond via RDKit, then complete the synthons into reactants by either leaving groups selection [27], graph generation [26], or SMILES generation [34,33]. In contrast, MEGAN [20] reframes the generative procedure as a sequence of graph edits that are completed by RDKit.…”

Section: Retrosynthesis Predictionmentioning

confidence: 99%

“…The third class is semi-template-based methods, which combine the advantages of both the generative models and the additional chemical knowledge. Conventional frameworks [34,26,27,33] in this category follow the same idea: They first identify the reactive bond and convert the product into synthons by RDKit. Then, another model completes synthons into reactants.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Retroformer: Pushing the Limits of Interpretable End-to-end Retrosynthesis Transformer

Wan¹,

Liao²,

Hsieh³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Retrosynthesis prediction is one of the fundamental challenges in organic synthesis. The task is to predict the reactants given a core product. With the advancement of machine learning, computer-aided synthesis planning has gained increasing interest. Numerous methods were proposed to solve this problem with different levels of dependency on additional chemical knowledge. In this paper, we propose Retroformer, a novel Transformer-based architecture for retrosynthesis prediction without relying on any cheminformatics tools for molecule editing. Via the proposed local attention head, the model can jointly encode the molecular sequence and graph, and efficiently exchange information between the local reactive region and the global reaction context. Retroformer reaches the new state-of-the-art accuracy for the end-to-end template-free retrosynthesis, and improves over many strong baselines on better molecule and reaction validity. In addition, its generative procedure is highly interpretable and controllable. Overall, Retroformer pushes the limits of the reaction reasoning ability of deep generative models.

show abstract

Computer‐Assisted Synthesis Planning

Tu,

Levin,

Coley

2023

Enabling Tools and Techniques for Organic Synthesis

View full text Add to dashboard Cite

RetroPrime: A Diverse, plausible and Transformer-based method for Single-Step retrosynthesis predictions

Cited by 70 publications

References 32 publications

Retrosynthetic reaction pathway prediction through neural machine translation of atomic environments

Retrosynthetic reaction pathway prediction through neural machine translation of atomic environments

Retroformer: Pushing the Limits of Interpretable End-to-end Retrosynthesis Transformer

Computer‐Assisted Synthesis Planning

Contact Info

Product

Resources

About