The Number of the Nearest Neighboring Sites in the Local Maximum Algorithm for Identifying Peaks in Mass Spectrometry Data Analysis

• Activity Number: 582
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Zhiying You*+ and Andrzej Kulczycki and Sreelatha Meleth and Denise Oelschlager and William Grizzle
• Companies: The University of Alabama at Birmingham and The University of Alabama at Birmingham and The University of Alabama at Birmingham and The University of Alabama at Birmingham and The University of Alabama at Birmingham
• Keywords: mass spectrometry data ; local maximum algorithm ; peak identification ; number of the nearest sites ; curvature

Abstract

The local maximum algorithm used to identify peaks in mass spectrometry data analysis looks for m/z values whose intensity is the highest among the nearest neighboring S sites. The value S impacts on the number of identified peaks but we lack for a criterion for the selection of value S. We propose to determine value for S by using a criterion of maximum curvature (MC) of the curve plotting the number of identified peaks. We repeat the local maximum algorithm and fit a curve equation to the number of identified peaks; we then calculate curvature and determine S such that the curvature obtains its maximum at S. This implies geometrically the curve changes its direction most at S. A different from S value may be used and we are able to character it, given the MC and value S. The prerequisite to the application of this approach is that the resulted equation has a non-zero second derivative.