Datasets and Their Influence on the Development of Computer Assisted Synthesis Planning Tools in the Pharmaceutical Domain

Thakkar, Amol; Kogej, Thierry; Reymod, Jean-Louis; Engkvist, Ola; Bjerrum, Esben Jannik

doi:10.26434/chemrxiv.9897692.v1

Cited by 3 publications

(3 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Of notable mention is that, comparing various data sources, including patents (USPTO and Pistachio), literature and patents (Reaxys), and industrial data (AstraZeneca ELN), despite similarities in their size of template sets, they differ in the coverage of reaction space. Reaxys stands out for its extensive and uniquely diverse collection of reaction templates, providing a broader reaction space [60].…”

Section: Data Sourcesmentioning

confidence: 99%

Application of Modern Intelligent Algorithms in Retrosynthesis Prediction

Liao,

Shi,

Zhu

2024

Preprint

View full text Add to dashboard Cite

In recent years, with the rapid advance of computer science, various modern intelligent algorithms have successively emerged. Transformer, based on multi-head attention mechanism, is one of the most favored AI models among in this century. The introduction of these algorithms leads to dramatic progress in retrosynthesis prediction. Unlike conventional retrosynthesis prediction models, retrosynthesis prediction based on intelligent algorithms can automatically extract chemistry knowledge from chemical reaction datasets to predict retrosynthesis routes. In this review, we provide a comprehensive overview of retrosynthesis prediction based on modern intelligent algorithms, particularly artificial intelligence algorithm. After introducing the related deep learning model, the existing chemical reaction datasets and molecular representations are presented. Subsequently, the current state-of-the art of AI-assisted retrosynthesis prediction models in recent years is discussed, including template-based models, template-free models, and semi-template-based models. Additionally, we conclude by comparing retrosynthesis prediction models across different categorizations. Finally, several challenges and limitations of these current methods are summarized, with a view to promising directions for future research.

show abstract

Section: Data Sourcesmentioning

confidence: 99%

Application of Modern Intelligent Algorithms in Retrosynthesis Prediction

Liao,

Shi,

Zhu

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Several methods have been proposed to predict synthetic pathways using machine learning. 49,50 We used AIZynthFinder to determine whether the optimized molecule has a practical synthetic pathway.…”

Section: ■ Experimental Sectionmentioning

confidence: 99%

Gargoyles: An Open Source Graph-Based Molecular Optimization Method Based on Deep Reinforcement Learning

Erikawa,

Yasuo,

Suzuki

et al. 2023

ACS Omega

View full text Add to dashboard Cite

Automatic optimization methods for compounds in the vast compound space are important for drug discovery and material design. Several machine learning-based molecular generative models for drug discovery have been proposed, but most of these methods generate compounds from scratch and are not suitable for exploring and optimizing user-defined compounds. In this study, we developed a compound optimization method based on molecular graphs using deep reinforcement learning. This method searches for compounds on a fragment-by-fragment basis and at high density by generating fragments to be added atom by atom. Experimental results confirmed that the quantum electrodynamics (QED), the optimization target set in this study, was enhanced by searching around the starting compound. As a use case, we successfully enhanced the activity of a compound by targeting dopamine receptor D2 (DRD2). This means that the generated compounds are not structurally dissimilar from the starting compounds, as well as increasing their activity, indicating that this method is suitable for optimizing molecules from a given compound. The source code is available at .

show abstract

“…1,[6][7][8][9][10] Despite efforts in building models that effectively learn chemistry from data, the quality of the data sets remains the primary limitation on performance improvements. The impact of data sets sizes and variability on the performance of computer-assisted synthesis planning tools has been recently investigated by Thakkar et al 11 . Nevertheless, the influence of having chemically wrong examples in training data sets remains a topic of little research regardless of its relevance and impact in all data-driven chemical applications.…”

Section: Introductionmentioning

confidence: 99%

Unassisted Noise-Reduction of Chemical Reactions Data Sets

Toniato

Schwaller²,

Cardinale³

et al. 2020

Preprint

View full text Add to dashboard Cite

<div><div><div><p>Existing deep learning models applied to reaction prediction in organic chemistry are able to reach extremely high levels of accuracy (> 90% for NLP- based ones1). With no chemical knowledge embedded than the information learnt from reaction data, the quality of the data sets plays a crucial role in the performance of the prediction models. While human curation is prohibitively expensive, the need for unaided approaches to remove chemically incorrect entries from existing data sets is essential to improve the performance of artificial intelligence models in synthetic chemistry tasks. Here we propose a machine learning-based, unassisted approach to remove chemically wrong entries (noise) from chemical reaction collections. Results show that models trained on cleaned and balanced data sets improve the quality of the predictions without a decrease in performance. For the retrosynthetic models the round-trip accuracy is enhanced by 13% and the value of the cumulative Jensen Shannon metric is lowered down to 70% of its original value, while maintaining high values of coverage (97%) and constant class-diversity (1.6) at inference.</p></div></div></div>

show abstract

Datasets and Their Influence on the Development of Computer Assisted Synthesis Planning Tools in the Pharmaceutical Domain

Cited by 3 publications

References 38 publications

Application of Modern Intelligent Algorithms in Retrosynthesis Prediction

Application of Modern Intelligent Algorithms in Retrosynthesis Prediction

Gargoyles: An Open Source Graph-Based Molecular Optimization Method Based on Deep Reinforcement Learning

Unassisted Noise-Reduction of Chemical Reactions Data Sets

Contact Info

Product

Resources

About