The BAMBOO method for correcting batch effects in high throughput proximity extension assays for proteomic studies.

Area of Science:

• Proteomics
• Biostatistics
• Bioinformatics

Background:

• Proximity extension assay (PEA) is crucial for large-scale proteomic studies.
• Batch effects, including protein-specific, sample-specific, and plate-wide variations, can compromise data integrity and lead to false discoveries.
• Existing methods for batch effect correction in PEA are not fully optimized.

Purpose of the Study:

• To characterize batch effects in PEA proteomics.
• To develop and validate a robust method for correcting batch effects using bridging controls.
• To enhance the reliability of large-scale PEA proteomic analyses.

Main Methods:

• Characterization of three types of batch effects in PEA data.
• Development of BAMBOO (Batch Adjustments using Bridging cOntrOls), a novel regression-based batch correction method.
• Comparative analysis of BAMBOO against median centering, median of the difference (MOD), and ComBat using simulations and experimental data.

Main Results:

• BAMBOO and MOD demonstrated robustness against outliers in bridging controls, unlike median centering and ComBat.
• Optimal batch correction was achieved with 10-12 bridging controls.
• BAMBOO and MOD significantly reduced the incidence of false discoveries compared to other methods in experimental validation.

Conclusions:

• Batch effects are prevalent in PEA proteomic studies and necessitate effective correction strategies.
• BAMBOO offers a robust and effective solution for mitigating batch effects in PEA.
• The findings advocate for the adoption of BAMBOO to improve the accuracy and reliability of large-scale proteomic analyses.