ProteoIntegrator is a scalable, web-based platform for harmonizing and analyzing multisource proteomics data, addressing missing values and batch effects that hinder large-scale data integration. We benchmarked 12 imputation methods on DDA/DIA and multi-instrument data sets, identifying Bayesian principal component analysis (BPCA) as optimal for integrating with parallelized ComBat for efficient batch correction capable of handling >1000 samples under 1 min. This no-code tool automates quality control (filtering, normalization, and imputation), differential expression analysis with interactive visualizations, and conducts pathway enrichment and protein–protein interaction network analysis. It outperforms HarmonizR, a nonimputation-based batch correction method, in the detection of differentially expressed proteins (DEPs). Demonstrations include correction of 7 year QC drifts across three instruments, integrating DDA/DIA/TMT data from varied platforms and gradients with >98% quantification accuracy, cross-laboratory analyses on five mass spectrometers, and harmonization of public data sets. Its utility is demonstrated in reanalyzing two published kidney cancer data sets, recapitulating metabolic shifts such as OXPHOS downregulation and glycolytic upregulation. The platform is publicly accessible at https://pi.meritudio.com to empower reproducible proteomics discovery.
Ren et al. (Sat,) studied this question.