Constrained Smoothing of Scatterplots with Applications to Mass Spectrometry Data

• Activity Number: 306
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Xingdong Feng*+ and Nell Sedransk
• Companies: National Institute of Statistical Sciences and National Institute of Statistical Sciences
• Keywords: Smoothing ; regression splines ; shape restrictions

Abstract

A certain concentration level of a given protein needs to be reached for a satisfactory precision of the measured signal intensities from the mass spectrometry systems in proteomic analysis, and linear regressions are usually used to calibrate the signal measurements and to estimate these levels. In the real data analysis, we also observe that the variations are larger in the boundaries of the concentration levels, so it may imply that not only the monotonicity is needed for the smoothing curve but also the convexity is needed for the variation pattern across the concentration levels. To account for both restrictions, we propose an iterated method based on existing smoothing methods. This hybrid method can achieve the global optimal rates under certain weak conditions.