Appeal 2007-0537 Application 10/102,902 residues for one or more of the 20 known proteins) “which correspond to features of a query molecule fragment pair” (chemical shifts and sequences for a triplet of amino acids residues for the unknown structure); and “aligning said candidate molecule fragment pair to said query molecule based on said corresponding features” (see FF 12 (citing Cornilescu 289 (abstract), 293 (Figure 1) (describing differences resulting from such alignment)); FF 23). Finally, with respect to the disputed claim language (see App. Br. 11; Reply Br. 1-4), Cornilescu’s “features of said candidate molecule fragment pair and query molecule fragment pair comprise descriptors” (mathematical representations of chemical shifts and amino acid sequences for computer analysis), which are weighted using “context-adaptive scaling of said descriptors.” (FFs 1-19, 22-29.) Appellants argue: In direct contrast to Cornilescu, the claimed invention recognizes the importance of fragment pairs. Indeed, an important feature of an exemplary aspect of the claimed invention may include partitioning a molecule into fragments, and using data regarding two of the fragments (e.g. neighboring fragments connected by a rotatable bond) as a basis for a similarity search. The Application explains that there may be significant benefits to using features of fragment pairs (e.g., neighboring fragments connected by a rotatable bond), as a basis for a similarity search. For example, the Application states that a “key” to the invention is that “it avoids the need to search every conformation of every molecule and apply a similarity metric to the query at each possible alignment. This is done by using selected features from fragment pairs to align them to the query” (Application at page 15, line 24-page 16, line 5). Nowhere is this recognized by Cornilescu. 12Page: Previous 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Next
Last modified: September 9, 2013