This work introduces GraphormerMapper – a new algorithm for reactions atom-to-atom mapping (AAM) based on a distance-aware BERT neural network. In benchmarking studies with IBM RxnMapper, the best AAM algorithm according to our previous study, we demonstrate that our AAM algorithm is superior on our “Golden” benchmarking dataset. The mapper is implemented in Chython [https://github.com/chython/chython] and Chytorch [https://github.com/chython/chytorch, https://github.com/chython/chytorch-rxnmap] Python packages which are freely available for out-the-box use. Chython is a cheminformatics library with a simple interface for processing reaction and molecular data. The key features of Chython are: chemical functional groups standardization, checking atom valence errors, substructure search, and advanced reaction manipulation, for example, generating products from reactants and reaction atom-to-atom mapping. Chytorch provides a PyTorch-like interface for graph-based neural networks developed specifically for chemical tasks.

Artificial Intelligence is revolutionizing many aspects of the pharmaceutical industry. Deep learning models are now routinely applied to guide drug discovery projects leading to faster and improved findings, but there are still many tasks with enormous unrealized potential. One such task is the reaction yield prediction. Every year more than one fifth of all synthesis attempts result in product yields which are either zero or too low. This equates to chemical and human resources being spent on activities which ultimately do not progress the programs, leading to a triple loss when accounting for the cost of opportunity in time wasted. In this work we pre-train a BERT model on more than 16 million reactions from 4 different data sources, and fine tune it to achieve an uncertainty calibrated global yield prediction model. This model is an improvement upon state of the art not just from the increase in pre-train data but also by introducing a new embedding layer which solves a few limitations of SMILES and enables integration of additional information such as equivalents and molecule role into the reaction encoding, the model is called BERT Enriched Embedding (BEE). The model is benchmarked on an open-source dataset against a state-of-the-art synthesis focused BERT showing a near 20-point improvement in r2 score. The model is fine-tuned and tested on an internal company data benchmark, and a prospective study shows that the application of the model can reduce the total number of negative reactions (yield under 5%) ran in Janssen by at least 34%. Lastly, we corroborate the previous results through experimental validation, by directly deploying the model in an on-going drug discovery project and showing that it can also be used successfully as a reagent recommender due to its fast inference speed and reliable confidence estimation, a critical feature for industry application.

This work introduces GraphormerMapper, a new algorithm for reaction atom-to-atom mapping (AAM) based on a transformer neural network adopted for the direct processing of molecular graphs as sets of atoms and bonds, as opposed to SMILES/SELFIES sequence-based approaches, in combination with the Bidirectional Encoder Representations from Transformers (BERT) network. The graph transformer serves to extract molecular features that are tied to atoms and bonds. The BERT network is used for chemical transformation learning. In a benchmarking study with IBM RxnMapper, which is the best AAM algorithm according to our previous study, we demonstrate that our AAM algorithm is superior to it on our "Golden" benchmarking data set.

The quality of experimental data for chemical reactions is a critical consideration for any reaction-driven study. However, the curation of reaction data has not been extensively discussed in the literature so far. Here, we suggest a 4 steps protocol that includes the curation of individual structures (reactants and products), chemical transformations, reaction conditions and endpoints. Its implementation in Python3 using CGRTools toolkit has been used to clean three popular reaction databases Reaxys, USPTO and Pistachio. The curated USPTO database is available in the GitHub repository (Laboratoire-de-Chemoinformatique/Reaction_Data_Cleaning).

In the context of drug discovery, a key problem is the identification of candidate molecules that affect proteins associated with diseases. Inside Janssen Pharmaceutica, the Chemogenomics project aims to derive new candidates from existing experiments through a set of machine learning predictor programs, written in single-node C++. These programs take a long time to run and are inherently parallel, but do not use multiple nodes. We show how we reimplemented the pipeline using Apache Spark, which enabled us to lift the existing programs to a multi-node cluster without making changes to the predictors. We have benchmarked our Spark pipeline against the original, which shows almost linear speedup up to 8 nodes. In addition, our pipeline generates fewer intermediate files while allowing easier checkpointing and monitoring.