Some Selected Research Accomplishments

• 2003 introduced an algorithm that assigns unique IDs to chemical compounds. In 2003, Kasturi, Hamelberg, Liu and I showed that there is a natural graph algorithm that can assign essentially unique IDs to chemical compounds. We call these unique chemical IDs or UCKs [40]. The basic idea was to define certain classes of natural operations on labeled graphs. For example, one can assign a new (long) label to a node by forming the set of all labeled paths of depth d or less originating from the node, lexigraphically ordering the resulting set, and then concatenating the results. Using a few such operations one can define keys for labeled graphs, which we showed are unique for the types of graphs associated with chemical compounds. In contrast, today, most databases of chemical compounds assign ascension numbers to new chemical compounds when they are added to the database (1, 2, 3, etc.). We showed that over 10% of the chemical compounds in one of the most popular databases of chemical compounds were duplicates and were not detected because of the defects in the current ascension numbers [40]. The lack of keys for chemical compounds makes it extremely difficult to compare chemical compounds across two different databases. Using UCKs, data can be integrated easily from multiple databases containing chemical compounds, facilitating the discovery of new drugs and therapeutic treatments.