Reproducibility of SELDI-TOF protein patterns in serum: comparing datasets from different experiments

摘要: Motivation: There has been much interest in using patterns derived from surface-enhanced laser desorption and ionization (SELDI) protein mass spectra serum to differentiate samples patients both with without disease. Such have used identification of the underlying proteins responsible. However, there are questions as stability this procedure over multiple experiments. Results: We compared SELDI proteomic three experiments by same group on separating ovarian cancer normal tissue. These available web at http://clinicalproteomics.steem.com. In general, results were not reproducible across experiments. Baseline correction prevents reproduction for two one experiment, is evidence a major shift protocol mid-experiment which could bias results. another, structure noise regions allows us distinguish cancer, suggesting that normals cancers processed differently. Sets features found discriminate well experiment do generalize other Finally, calibration all appears suspect. Taken together, these concerns suggest uncovered be due artifacts sample processing, biology cancer. provide some guidelines design analysis like ensure better reproducible, biologically meaningfully results. Availability: The MATLAB Perl code our analyses http://bioinformatics.mdanderson.org