Enhanced differential expression statistics for data-independent acquisition proteomics

Tomi Suomi,Laura L Elo

doi:10.1038/s41598-017-05949-y

Abstract

We describe a new reproducibility-optimization method ROPECA for statistical analysis of proteomics data with a specific focus on the emerging data-independent acquisition (DIA) mass spectrometry technology. ROPECA optimizes the reproducibility of statistical testing on peptide-level and aggregates the peptide-level changes to determine differential protein-level expression. Using a ‘gold standard’ spike-in data and a hybrid proteome benchmark data we show the competitive performance of ROPECA over conventional protein-based analysis as well as state-of-the-art peptide-based tools especially in DIA data with consistent peptide measurements. Furthermore, we also demonstrate the improved accuracy of our method in clinical studies using proteomics data from a longitudinal human twin study.

Highlights

The onset of high-throughput technology has enabled us to quantify complex protein mixtures using mass spectrometers
An emerging technology, called data-independent acquisition (DIA), capitalizes on the strengths of both the shotgun and targeted methods by combining the reproducibility of selected reaction monitoring (SRM) with the extensive number of proteins identified in shotgun proteomics[4,5,6]
We first tested the applicability of our Reproducibility-optimized peptide change averaging (ROPECA) method using the DIA profiling standard benchmark data, which contains 12 non-human proteins spiked into a constant human background in different concentrations[1] (Table 1)

Summary

Introduction

The onset of high-throughput technology has enabled us to quantify complex protein mixtures using mass spectrometers. A widely used option for obtaining a global protein profile for a sample is label-free shotgun proteomics where the mass spectrometer is operated in data-dependent acquisition (DDA) mode In this technique, the most intense precursor ions from a survey scan are isolated and fragmented to produce tandem mass spectra (MS/MS or MS2), which are matched against a database of known sequences for peptide identification. This approach uses the capabilities of triple quadrupole mass spectrometers to filter and selectively monitor a specific molecular ion and their corresponding fragment ions generated by collisional dissociation These precursor-fragment ion pairs, termed SRM transitions, are repeatedly measured and counted over time, enabling reproducible quantification of the target peptides[2, 3]. ROPECA first optimizes the reproducibility of statistical testing for each data separately by maximizing the overlap of www.nature.com/scientificreports/

Objectives

Methods

Results