Empirical Testing of New Tools for Searching Generic (Markush) Structures in Main Chemistry Databases

Abstract

Generic or Markush structures in chemistry patents represent a very useful and convenient description of compounds claimed in patent documents. On the other hand, chemists face an uncertainty about a real existence of chemical compounds represented by a given Markush structure. Particularly for a synthetic chemist, generic or Markush structure brings uncertainty about which specific compound corresponding to a generic formula was really synthesized and identified and which not (and is included for the maximum extent of patent claims). This problem should in first instance be solved by chemists creating bibliographic records for chemistry databases, primarily SciFinder and Reaxys. In case of a really synthesized compound, it is included in the existing list of known chemical compounds and receives the so called Registry number, most importantly the CAS RN. SciFinder database offers a chemist some help by setting up a new category of compounds with only partial description under the name Prophetic in Patents. Markush formulas represent a serious problem when we search for particular chemical compounds. Both chemical databases, SciFinder and Reaxys, recently offered new searching tools which make searching for generic compounds possible. The aim of this work is to test these searching tools on a random set of chemistry patents which contain both Markush formulas and compounds in the Prophetic category. Despite the limited set of patents tested, it was possible to formulate some conclusions about the applicability of these tools. Particularly the category Prophetic in Patents seems to be useful, as well as individual registration of Markush structures with relevant Reaxys Registry Number.