Simultaneous and exact interval estimates for the contrast of two groups based on an extremely high dimensional...

Area of Science:

• Biostatistics
• Bioinformatics
• Proteomics
• Genomics

Background:

• High-throughput proteomic and genomic data analysis often relies on qualitative P-values for biomarker identification.
• Quantitative interval estimation for group contrasts offers a more informative approach for biomarker discovery.
• Existing methods for analyzing surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS) and microarray data have limitations in providing magnitude of differences.

Purpose of the Study:

• To develop a simultaneous confidence bands method for identifying potential biomarkers in high-dimensional datasets.
• To provide a quantitative measure of differences between treatment groups, complementing traditional P-value based methods.
• To control the overall confidence coverage level in biomarker detection.

Main Methods:

• A permutation scheme was employed to construct simultaneous confidence bands for mean differences between two groups.
• The method analyzes entire spectra in SELDI-TOF MS data and constructs confidence bands for mean differences.
• Peaks were identified based on maximal differences determined by the confidence bands, providing both qualitative and quantitative data.

Main Results:

• The confidence bands method successfully identified potential biomarkers in ovarian cancer SELDI-TOF MS data and spiked-in protein samples.
• Analysis revealed minimal absolute differences for some previously identified ovarian cancer biomarkers, while also detecting new markers with greater intensity differences.
• The method accurately detected spiked-in peaks, adducts, and double-charged species, and was applied to prostate cancer microarray data for fold change analysis.

Conclusions:

• The developed confidence bands method offers a robust, quantitative approach for biomarker discovery in high-dimensional omics data.
• This method enhances the interpretation of biomarker significance by providing the magnitude of differences between groups.
• The approach is versatile, applicable to both SELDI-TOF MS and microarray data for identifying and validating potential biomarkers.