Network Analysis of the Organic Chemistry in Patents, Literature, and Pharmaceutical Industry

Chemical reactions can be connected in large networks such as knowledge graphs. In this way, prior work has been able to draw meaningful conclusions about the structures and properties of the included organic chemistry. However, the research has focused on public sources of organic chemistry that might lack the intricate details of the synthesis routes used in in-house drug discovery. In this work, we expand on previous analyses to also include an in-house electronic lab notebook (ELN), such that important differences between the network architectures can be investigated. Three chemical reaction knowledge graphs were constructed from US Patent and Trademark Office (USPTO), Reaxys, and an in-house ELN, respectively. The three knowledge graphs were compared. We found that the Reaxys knowledge graph is the most interconnected, whereas the USPTO and ELN knowledge graphs appear more arranged around a few central nodes. These differences might be attributed to the different origins of the data in the three sources.