A statistical framework for combining and interpreting proteomic datasets

Area of Science:

• Proteomics
• Computational Biology
• Statistical Modeling

Background:

• Accurate protein function identification requires integrating high-throughput proteomic data.
• High-throughput datasets often contain errors, leading to incomplete and contradictory information.
• Existing methods struggle with error rates and combining diverse experimental results.

Purpose of the Study:

• To develop a statistical framework for interpreting and integrating high-throughput proteomic data.
• To accurately calculate the probability of proteins belonging to the same complex.
• To combine information from multiple experiments to enhance coverage and accuracy.

Main Methods:

• Development of a statistical framework based on Bayes' law.
• Application to two protein complex purification datasets.
• Estimation of false positive and false negative error rates without reference sets.
• Technique for estimating the detectable proteome size.

Main Results:

• Accurate calculation of pairwise protein complex probabilities using high-throughput data.
• Successful integration of two datasets, yielding improved coverage and accuracy.
• Method does not require a priori verified protein interaction data.
• Provides a technique for estimating the total number of detectable proteins.

Conclusions:

• The developed statistical framework effectively interprets and integrates noisy high-throughput proteomic data.
• The approach enhances the accuracy and coverage of protein-protein interaction networks.
• The method offers a robust way to assess experimental technique capabilities.