In the last two decades, mass spectrometry-based workflows have emerged as the preferred approaches for proteomic analysis, however, there are some concerns regarding its quantitative capabilities in discovery proteomic studies, due to the large amount of missing values across data sets.
Solving the Missing Value Problem with DDA
In order to solve this issue, scientists are converging on a new way to process and score signals from MS1 Data-Dependent Acquisition (DDA) to resolve the “missing value” problem. Compared with traditional workflows for DDA, these new software scoring techniques achieve deeper proteome coverage, fewer missing values, and lower quantification variance. The method also enables flexible and robust proteome characterization based on covariation of peptide abundances.1
Label-Free Quantification (LFQ) is one of the most efficient approaches for quantifying proteome differences using mass spectrometry. It is cost effective, takes less time, and is less process-intensive than labelled methods. DDA and Data Independent Acquisition (DIA) are the two methods of acquiring information in proteomics using LFQ-dependent mass spectrometry. DDA has been the preferred method for years, but is plagued by the stochastic nature of precursor selection and low sampling efficiency due to the limited speed of mass spectrometers. These challenges result in missing individual peptide identification across LC-MS/MS runs within a larger dataset—even after measurements are replicated. This is known as the missing value problem.
The problem significantly limits the size of the DDA-acquired proteomics dataset across which reliable quantification can be made for each protein. As a result, scientists created DIA methods to overcome the limitations of DDA and the missing value issue, however, DDA remains the gold standard due to its ability to detect a wider dynamic range of proteins in a complex matrix. DDA would undoubtedly be preferred if the problem of the missing values could be solved.
Maximizing Reproducible Protein Quantitation
Scientific teams from around the world are harnessing the increased computing power and storage available to introduce a new layer of digital quality testing in DDA workflows. At the Karolinska Institute in Sweden, Professor Roman Zubarev, proposed a new, improved quantification-centric approach to DDA, improving several features of the traditional identification-centric approach to signal processing. 1 The team reasons that missing values are not intrinsic to DDA approaches since the signal from the molecular ion is usually present among the MS1 spectra. The signal information simply needs to be processed differently, with less focus on identification prior to quantification. This new analytical workflow recovers missing values using a protein scoring scheme for quality control.
Like what you are learning?
In Germany, Dr. Kirti Sharma and her team at the Max Planck Institute of Biochemistry investigated the mouse brain proteome—by cell type and brain region—using a similar DDA method but combined it with deep sequencing-based transcriptome analysis to fully map transcripts and protein expression. The result is a deep proteomic profile of the mouse brain that can serve as a rich resource for brain development and functional analyses on a systems level.2
In the United States, scientists led by Dr. Daniel Lopez-Ferrer at Thermo Fisher Scientific (San Jose, California) have demonstrated the value of LFQ when combining DDA with a new and improved version of Thermo Scientific™ Proteome Discoverer™ software. This newly updated software includes a novel processing step that re-calibrates retention times and mass deviations, as well as searches for features in the MS1 scans, the scan where all of the sample compounds are recorded. The abundance of these compounds (peptides) is then linked to their corresponding identifications across the various runs, resulting in extremely precise profile of the proteome. Dr. Lopez-Ferrer’s team’s data shows that the number of missing values remains below 5%. Additionally, the median variation of the abundance is lower than in matching DIA datasets, and that the number of quantified proteins is 25% higher. These results provide additional confirmation that MS1-DDA is the best technology for performing LFQ in quantitative proteomics.
“DDA remains the preferred technology for deep and accurate proteomic investigations,” said Lopez-Ferrer. “The newest developments in the Proteome Discoverer platform, in combination with the top-of-the-line Thermo Scientific™ Q-Exactive™ HF Mass Spectrometer using DDA methodology together provide a solution for precise data acquisition and analysis for a broad spectrum of applications in quantitative proteomics.”
Scientists from each of the three teams will be presenting at ASMS 2016 in a workshop titled Digital Proteome Maps: Label free Protein Quantification and DDA.
Learn more at ThermoScientific.com/ASMS
- Zhang, B., et al. (2016) “DeMix-Q: Quantification-centered data processing workflow,” Molecular & Cellular Proteomics, 15(10) (pp. 1467–1478). mcp.O115.055475
- Sharma, K. (2015) “Cell type—and brain region—resolved mouse brain proteome,” Nature Neuroscience, 18(12) (pp. 1819–1831).