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

Area of Science:

• Biochemistry
• Proteomics
• Biomarker Discovery

Background:

• Surface-enhanced laser desorption and ionization (SELDI) protein mass spectra show promise for disease detection.
• Identifying disease-specific protein patterns from serum is of significant interest.
• Concerns exist regarding the reproducibility of SELDI-based diagnostic procedures.

Purpose of the Study:

• To evaluate the reproducibility of SELDI proteomic spectra across multiple experiments for ovarian cancer detection.
• To identify potential sources of variability and artifacts in SELDI serum analysis.
• To provide guidelines for improving the reliability of SELDI-based biomarker studies.

Main Methods:

• Comparison of SELDI proteomic spectra from three independent serum experiments.
• Analysis of spectral data for reproducibility and potential artifacts.
• Evaluation of baseline correction and mass calibration methods.

Main Results:

• Significant irreproducibility of SELDI spectra was observed across the three experiments.
• Baseline correction failed to ensure consistent results in two experiments.
• Potential protocol shifts and processing differences between sample groups were identified.
• Feature sets effective in one experiment did not generalize to others.
• Suspect mass calibration across all experiments raises concerns.

Conclusions:

• Observed patterns in SELDI serum spectra may result from experimental artifacts rather than true biological differences.
• The reliability of SELDI for distinguishing cancer from normal tissue in these experiments is questionable.
• Recommendations are provided for designing more robust and reproducible SELDI proteomic studies.