The Chemical Abstract's Service generic chemical (Markush) structure storage and retrieval capability. 1. Basic concepts

The Concept of Novel Compositions of Matter: A Theoretical Analysis

“…It was then noted that more exact graph-theoretic similarity tests were being pushed further and further down the workflow and still achieving overall utility. Many other workers have had similar experiences, or expressed similar ideas in expectation of such experiences, though not usually or necessarily concerning regular expressions specifically; [22][23][24][25][26][27][28][29][30][31][32] for an example bibliography of earlier work, and particularly a study on "matching-relations, user-defined match levels, and transition from the reduced graph search to the refined search" [32]. This idea of increasing levels of rigor of testing which catch more and more compounds as in some sense similar, when at first examination they are not so, is important in exploring the notion of similarity.…”

“…This idea of increasing levels of rigor of testing which catch more and more compounds as in some sense similar, when at first examination they are not so, is important in exploring the notion of similarity. However, in order to cover a set of molecules which the assignees sees as essentially similar, patents themselves contain the above-mentioned generalized descriptions of novel compositions of matter called Markush representations [6,22,23]. Searching them to see if a composition of matter is novel is a non-trivial task.…”

Mapping Chemical Structures to Markush Structures Using SMIRKS

“…There are several sophisticated commercial platforms for patent Markush database searching. [4][5][6][7][8] The search usually requires an initial screening of proprietary Markush structure databases (based on fragments and reduced graphs) and a subsequent atomby-atom search that is based on the Ullmann algorithm for subgraph isomorphism. [9,10] In contrast, our algorithm is straightforward with a broad software compliance by using SMIRKS.…”

Intellectual property information: A comparative analysis of main information providers

“…Searching for citations and references for whole patent families is possible only in Questel-Orbit databases. Chemical structure searching (including reactions) is possible on all commercial suppliers; searching Markush structures is possible only on QuestelOrbit (Borne, O'Hara, Roesch, & Skippon, 2003) and STN International (Ebe, Sanderson, & Wilson, 1991;Fisanick, 1990). DIALOG, EPO, and Questel-Orbit lack a database with gene sequences such as Derwent's GENESEQ, GenBank, or the sequence information in the Registry File of CAS and fuzzy search options, as presented within the context of STN International by, for example, Derwent's gene sequence database and the search programs BLAST and FASTA (Stock & Stock, 2003d, pp.…”