Reproducibility of proteomic profiles over 3 years in postmenopausal women not taking postmenopausal hormones

Abstract

Most proteomics studies examine one blood specimen per participant; however, it is unknown how well measures at one time point reflect an individual's long-term proteome pattern. Therefore, we examined the stability of the proteome over 3 years in postmenopausal women not taking hormones for at least 3 months using surface-enhanced laser desorption and ionization time-of-flight mass spectrometry. Using the Nurses' Health Study blood cohort, we randomly selected 60 women from a subset providing 2 to 3 blood samples over 3 years. Four different protein chip surfaces/plasma fractions were examined: unfractionated plasma on a CM10 and H50 chip, pH ≥ 9, plasma fraction on a CM10 chip, and the organic fraction on the H50 chip, all with a low- and high-energy transfer protocol. Participant and quality control samples were aligned to a reference sample and then peak intensity was assessed for all peaks identified in the reference sample. The average coefficient of variation (CV) of the peak intensity within conditions ranged from 16% (H50, organic, low protocol) to 63% (CM10, pH ≥ 9, high protocol). Generally, the CV and mean peak intensity of the quality control samples were inversely correlated (median -0.48). The mean intraclass correlation (ICC) within conditions ranged from 0.37 (H50, unfractionated, low protocol) to 0.68 (CM10, unfractionated, high protocol). For a signal-to-noise cutoff of 2.0, we observed 334 peaks, of which 241 (72%) had an ICC of ≥0.40. Although we observed a large range of CVs and ICCs, sufficient numbers of peaks had reasonable ICCs to suggest that protein peak reproducibility over 3 years was reasonable among postmenopausal women not taking hormones. Copyright © 2008 American Association for Cancer Research.