Appeal No. 1996-2240
Application No. 08/105,482
establishing a prima facie case of obviousness in the first instance has not been met.
Again, in setting forth the rejection in the first instance, the examiner does not address
appellant’s claim limitation respecting a frequency distribution data base. Nor are we
persuaded by the examiner’s treatment of this issue in responding to appellant’s
arguments. It is apparent from the specification that “conventional mathematical models”
are not interchangeable in the claimed method. Nevertheless, the examiner has not
explained why frequency distribution probability analysis would have been selected over
other models, such as linear regression analysis, stepwise regression analysis, or cluster
analysis, which cannot successfully distinguish between disease and non-disease
populations. See pages 19 through 21 of the Specification.
In our judgment, the reason advanced by the examiner for using frequency
distribution analysis in the claimed screening method (“. . . since the frequency distribution
shows distinct classifiable differences . . . “) stems from appellant’s
description in the specification, and not from the prior art. Accordingly, the rejection of
claims 1 through 32 under 35 U.S.C. § 103 as unpatentable over Miyagi, Long and the
admitted state of the prior art is reversed.
11
Page: Previous 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Next
Last modified: November 3, 2007