‘Ring Breaker’: Neural Network Driven Synthesis Prediction of the Ring System Chemical Space

Thakkar, Amol; Selmi, Nidhal; Reymond, Jean‐Louis; Engkvist, Ola; Bjerrum, Esben Jannik

doi:10.26434/chemrxiv.9938969

Cited by 8 publications

(10 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In fact, it is a common practice that templates with very few examples are excluded from the set of templates used for model training, due to the inability of neural networks to learn meaningful relationships between their input and these rare templates. 2,15 While it may be possible to augment training examples with more experimental data (increasing the ability to perform goal number two described above), or design models tasked to perform very specific types of retrosynthetic predictions (such as ring breaking 16 ), it is also possible and perhaps easier to augment the information about which templates could be relevant in silico (increasing the ability to perform goal number one described above). The augmentation of training examples for machine learning applications has been shown to improve model performance across many domains including image classification, 17 acoustic modeling, 18 natural language processing, 19 and has also been used to create chemically identical but syntactically different SMILES strings for cheminformatics applications.…”

Section: ■ Introductionmentioning

confidence: 99%

Data Augmentation and Pretraining for Template-Based Retrosynthetic Prediction in Computer-Aided Synthesis Planning

Fortunato

Coley

Barnes

et al. 2020

J. Chem. Inf. Model.

View full text Add to dashboard Cite

This work presents efforts to augment the performance of data-driven machine learning algorithms for reaction template recommendation used in computer-aided synthesis planning software. Often, machine learning models designed to perform the task of prioritizing reaction templates or molecular transformations are focused on reporting high accuracy metrics for the one-to-one mapping of product molecules in reaction databases to the template extracted from the recorded reaction. The available templates that get selected for inclusion in these machine learning models have been previously limited to those that appear frequently in the reaction databases and exclude potentially useful transformations. By augmenting open-access datasets of organic reactions with artificially calculated template applicability and pretraining a template relevance neural network on this augmented applicability dataset, we report an increase in the template applicability recall and an increase in the diversity of predicted precursors. The augmentation and pretraining effectively teaches the neural network an increased set of templates that could theoretically lead to successful reactions for a given target. Even on a small dataset of well curated reactions, the data augmentation and pretraining methods resulted in an increase in top-1 accuracy, especially for rare templates, indicating these strategies can be very useful for small datasets. File list (2)download file view on ChemRxiv Data_Augmentation_and_Pre_Training_for_Template_... (563.68 KiB) download file view on ChemRxiv SI_Data_Augmentation_and_Pre_Training_for_Templat...

show abstract

Section: ■ Introductionmentioning

confidence: 99%

Data Augmentation and Pretraining for Template-Based Retrosynthetic Prediction in Computer-Aided Synthesis Planning

Fortunato

Coley

Barnes

et al. 2020

J. Chem. Inf. Model.

View full text Add to dashboard Cite

show abstract

“…The application of text-mining techniques revealed more than 3.75 million records. This data set became a very popular source of reaction data, and several projects have already used this database to train models [12][13][14][15][16][17] . However, little attention was paid so far to the quality of the data representation in this dataset.…”

Section: Introductionmentioning

confidence: 99%

Atom-to-Atom Mapping: A Benchmarking Study of Popular Mapping Algorithms and Consensus Strategies

Madzhidov¹,

Lin²,

Nugmanov³

et al. 2020

Preprint

View full text Add to dashboard Cite

Here, we discuss a reaction standardization protocol followed by a comparison of popular Atom-to-atom mapping (AAM) tools (ChemAxon, Indigo, RDTool, NextMove and RXNMapper) as well as some consensus AAM strategies. For this purpose, a dataset of 1851 manually curated and mapped reactions was prepared (the Golden dataset) and used as a reference set. It has been found that RXNMapper possesses the highest accuracy, despite the fact that it has some clear disadvantages. Finally, RXNMapper was selected as the best tool, and it was applied to map the USPTO dataset. The standardization protocol used to prepare the data, as well as the data itself are available in the GitHub repository https://github.com/Laboratoire-de-Chemoinformatique.<br><br><br><br>

show abstract

“…look up similar reactions. Finally, we are working on improving the recommendation policy, by for instance utilizing the “ring breaker” policy [ 45 ]. All such extensions should be possible to implement easily in the current codebase because it has low complexity and Halstead effort.…”

Section: Future Developmentsmentioning

confidence: 99%

AiZynthFinder: a fast, robust and flexible open-source software for retrosynthetic planning

et al. 2020

Self Cite

View full text Add to dashboard Cite

We present the open-source AiZynthFinder software that can be readily used in retrosynthetic planning. The algorithm is based on a Monte Carlo tree search that recursively breaks down a molecule to purchasable precursors. The tree search is guided by an artificial neural network policy that suggests possible precursors by utilizing a library of known reaction templates. The software is fast and can typically find a solution in less than 10 s and perform a complete search in less than 1 min. Moreover, the development of the code was guided by a range of software engineering principles such as automatic testing, system design and continuous integration leading to robust software with high maintainability. Finally, the software is well documented to make it suitable for beginners. The software is available at http://www.github.com/MolecularAI/aizynthfinder.

show abstract

‘Ring Breaker’: Neural Network Driven Synthesis Prediction of the Ring System Chemical Space

Cited by 8 publications

References 22 publications

Data Augmentation and Pretraining for Template-Based Retrosynthetic Prediction in Computer-Aided Synthesis Planning

Data Augmentation and Pretraining for Template-Based Retrosynthetic Prediction in Computer-Aided Synthesis Planning

Atom-to-Atom Mapping: A Benchmarking Study of Popular Mapping Algorithms and Consensus Strategies

AiZynthFinder: a fast, robust and flexible open-source software for retrosynthetic planning

Contact Info

Product

Resources

About