101. Improving label-free quantitative proteomics strategies by distributing shared peptides and stabilizing variance
- Author
-
Laurence Florens, Michael P. Washburn, Zhihui Wen, and Ying Zhang
- Subjects
Normalization (statistics) ,Proteomics ,Chromatography ,Chemistry ,Quantitative proteomics ,Variance (accounting) ,Analytical Chemistry ,Intensity (physics) ,Standard curve ,Linear regression ,Biological system ,Peptides ,Smoothing ,Label free - Abstract
In a previous study, we demonstrated that spectral counts-based label-free proteomic quantitation could be improved by distributing peptides shared between multiple proteins. Here, we compare four quantitative proteomic approaches, namely, the normalized spectral abundance factor (NSAF), the normalized area abundance factor (NAAF), normalized parent ion intensity abundance factor (NIAF), and the normalized fragment ion intensity abundance factor (NFAF). We demonstrate that label-free proteomic quantitation methods based on chromatographic peak area (NAAF), parent ion intensity in MS1 (NIAF), and fragment ion intensity (NFAF) are also improved when shared peptides are distributed on the basis of peptides unique to each isoform. To stabilize the variance inherent to label-free proteomic quantitation data sets, we use cyclic-locally weighted scatter plot smoothing (LOWESS) and linear regression normalization (LRN). Again, all four methods are improved when cyclic-LOWESS and LRN are applied to reduce variation. Finally, we demonstrate that absolute quantitative values may be derived from label-free parameters such as spectral counts, chromatographic peak area, and ion intensity when using spiked-in proteins of known amounts to generate standard curves.
- Published
- 2015