Prediction of Optimal Conditions of Hydrogenation Reaction Using the Likelihood Ranking Approach

Abstract

The selection of experimental conditions leading to a reasonable yield is an important and essential element for the automated development of a synthesis plan and the subsequent synthesis of the target compound. The classical QSPR approach, requiring one-to-one correspondence between chemical structure and a target property, can be used for optimal reaction conditions prediction only on a limited scale when only one condition component (e.g., catalyst or solvent) is considered. However, a particular reaction can proceed under several different conditions. In this paper, we describe the Likelihood Ranking Model representing an artificial neural network that outputs a list of different conditions ranked according to their suitability to a given chemical transformation. Benchmarking calculations demonstrated that our model outperformed some popular approaches to the theoretical assessment of reaction conditions, such as k Nearest Neighbors, and a recurrent artificial neural network performance prediction of condition components (reagents, solvents, catalysts, and temperature). The ability of the Likelihood Ranking model trained on a hydrogenation reactions dataset, (~42,000 reactions) from Reaxys® database, to propose conditions that led to the desired product was validated experimentally on a set of three reactions with rich selectivity issues.

Generative Modeling to Predict Multiple Suitable Conditions for Chemical Reactions

“…We first evaluate each reaction condition to determine if the reaction elements of the necessary categories are well specified and all are in-stock or can be purchased. We then sort the reaction conditions in order of highest predicted reaction yield if a yield prediction model is available for use, , or, in order of highest probability score, i.e. , p θ ( ĉ * | R * → P * , z * ) .…”

“…We first evaluate each reaction condition to determine if the reaction elements of the necessary categories are well specified and all are in-stock or can be purchased. We then sort the reaction conditions in order of highest predicted reaction yield if a yield prediction model is available for use, 21,22 or, in order of highest probability score, 13 i.e., . It should be noted that a higher probability score of a reaction condition does not guarantee a higher reaction yield but indicates that the reaction condition is more likely to occur in the dataset.…”

“…They represented a chemical reaction using the Morgan circular fingerprint and built neural networks to sequentially predict the catalyst, solvent, reagent, and temperature required for a given reaction. Afonina et al built a neural network that takes the ISIDA fragment descriptor of a reaction as input to predict the temperature, pressure, additive, and catalyst as the reaction condition. Walker et al used the MACCS key fingerprint and evaluated various prediction models for solvent prediction.…”

“…Model. XXXX, XXX, XXX−XXX order of highest predicted reaction yield if a yield prediction model is available for use,21,22 or, in order of highest probability score,13 i.e.,…”

Exploring the Optimal Alloy for Nitrogen Activation by Combining Bayesian Optimization with Density Functional Theory Calculations

“…The optimal value is proposed based on an acquisition function, which is a function that takes into account the predicted value (mean) and predicted variance obtained by Gaussian process regression. , Various acquisition functions have been designed to propose globally optimal values by BO . In materials science, BO is used to search for compositions and structures that can endow certain optimal physical properties. − Furthermore, BO is widely used in chemistry to optimize catalysts and synthetic conditions. − In recent years, not only BO but also various informatics methods have been introduced in the fields of catalysis and reactivity prediction. − …”