Automatic mapping of atoms across both simple and complex chemical reactions

Abstract

Mapping atoms across chemical reactions is important for substructure searches, automatic extraction of reaction rules, identification of metabolic pathways, and more. Unfortunately, the existing mapping algorithms can deal adequately only with relatively simple reactions but not those in which expert chemists would benefit from computer’s help. Here we report how a combination of algorithmics and expert chemical knowledge significantly improves the performance of atom mapping, allowing the machine to deal with even the most mechanistically complex chemical and biochemical transformations. The key feature of our approach is the use of few but judiciously chosen reaction templates that are used to generate plausible “intermediate” atom assignments which then guide a graph-theoretical algorithm towards the chemically correct isomorphic mappings. The algorithm performs significantly better than the available state-of-the-art reaction mappers, suggesting its uses in database curation, mechanism assignments, and – above all – machine extraction of reaction rules underlying modern synthesis-planning programs.

DRACON: disconnected graph neural network for atom mapping in chemical reactions

“…The final model achieves 61% (see Table 3) full-match accuracy on detection of the main product and 60% on detection centers in the reaction. We can not compare the result with the previous works 8, 25,32 because our research focuses on finding atoms of the main product and centers of reactions. We have not built a full solution for the entire atom mapping problem or outcome prediction problem.…”

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

“…Ambiguous reactions and reactions with unclear mechanism were excluded from the dataset. In the process of Jaworski et al [23] data curation, 5 records were not interpreted as chemical reactions, 10 reactions were identified as duplicate d (prior to addition of 469 USPTO reactions), 30 reactions were discarded since they were comprised by the molecules with invalid valence s at some atoms (see an example in Figure 4), and 350 reactions contained radicals. group as a part of a structure is depicted in Figure 5), and around 350 reactions were restored: radicals were removed (we assume that it was an attempt of Jaworski and others to balance the reactions using single atoms in the form of radicals).…”

“…Several benchmarking studies on the performance of different AAM tools have already been carried out [21][22][23][24] .…”

“…However, only few publications reported comparisons of tools using manually mapped dataset as a ground truth [15,23] .…”

Transformer: Linking Atom Mapping and Neural Machine Translation

“…To automatically mapping corresponding atoms across both simple and complex chemical reactions, Jaworski et al reported a combined approach called Mappet with graph-theoretical considerations and optimization programs. [11] In this work, the chemical rules are introduced to atoms matching and the performance of the algorithm is superior to other methods. To sum up, AM is usually considered as graph matching or MILP in prior work.…”