APEX

APEX computes absolute protein abundances from standard liquid chromatography-tandem mass spectrometry (LC-MS/MS) proteomics data to provide quantitative measurements of protein expression across biological samples.

Key Features:

• Java implementation: Distributed as Java-based software for processing LC-MS/MS proteomics datasets.
• Machine learning correction: Employs machine learning algorithms to correct for variable peptide detection driven by peptide physicochemical properties.
• Sampling depth adjustment: Calculates each protein's sampling depth (observed peptide count) and adjusts it using learned probabilities of peptide identification.
• Wide dynamic range: Reports absolute protein concentrations across approximately three to four orders of magnitude.
• Validation with external measures: Shows agreement with control measurements and known correlations with mRNA abundances and codon bias.

Scientific Applications:

• Model organism proteomics: Applied to large-scale protein expression studies in Saccharomyces cerevisiae and Escherichia coli to quantify absolute protein levels.
• Quantifying regulatory contributions: Used to estimate transcriptional versus translational contributions to protein abundance, reporting that mRNA abundance explains ~73% of protein variance in yeast and ~47% in E. coli.

Methodology:

Calculates observed peptide counts per protein (sampling depth) and adjusts those counts by applying learned peptide detection probabilities via machine learning to correct for LC-MS/MS detection biases.