Statistical protein quantification and significance analysis in label-free LC-MS experiments with complex designs

Abstract

BackgroundLiquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) is widely used for quantitative proteomic investigations. The typical output of such studies is a list of identified and quantified peptides. The biological and clinical interest is, however, usually focused on quantitative conclusions at the protein level. Furthermore, many investigations ask complex biological questions by studying multiple interrelated experimental conditions. Therefore, there is a need in the field for generic statistical models to quantify protein levels even in complex study designs.ResultsWe propose a general statistical modeling approach for protein quantification in arbitrary complex experimental designs, such as time course studies, or those involving multiple experimental factors. The approach summarizes the quantitative experimental information from all the features and all the conditions that pertain to a protein. It enables both protein significance analysis between conditions, and protein quantification in individual samples or conditions. We implement the approach in an open-source R-based software package suitable for researchers with a limited statistics and programming background.ConclusionsWe demonstrate, using as examples two experimental investigations with complex designs, that a simultaneous statistical modeling of all the relevant features and conditions yields a higher sensitivity of protein significance analysis and a higher accuracy of protein quantification as compared to commonly employed alternatives. The software is available at http://www.stat.purdue.edu/~ovitek/Software.html.

Phosphoproteomic characterization of DNA damage response in melanoma cells following MEK/PI3K dual inhibition

“…To identify proteins most significantly affected by the stimulus and to address the challenges posed by missing values, we used linear mixed-effects modeling (LiME). LiME, an improvement over ad hoc cutoffs or simple feature averaging, takes advantage of inherent replicate structure of the data and leverages information from a series of biological conditions to identify the significantly affected proteins (26,27). For PRKDC, LiME analysis revealed a systematic increase in peak area for the [s/t]Q containing phosphoPSMs, with >100-fold increase (P < 0.001) between combo and control treatments (Fig.…”

“…For MS studies, phosphopeptides were captured using [s/t]Q phosphomotif antibodies using PTMscan protocols (Cell Signaling Technology) (14) and were analyzed on an Orbitrap XL or Orbitrap-Velos. Database searches were performed using Mascot, peak areas determined using VistaQuant (25), and protein-level effects were assessed using LiME analysis (27). A detailed overview of the methods used here is presented in SI Materials and Methods and in Datasets S1-S3.…”

DeMix-Q: Quantification-Centered Data Processing Workflow

“…An unfortunate consequence of such a miss is that sometimes MS/MS-inferred peptides with vastly deviating abundances in run-torun or sample-to-sample comparisons are attributed to the same protein. These deviating peptides with questionable identities can drastically worsen the variances of protein abundances in the whole dataset and thus reduce the statistical power of the experiment (31). In contrast, in a quantification-centered approach, peptide abundance is the central factor to be investigated (9,15,(32)(33)(34).…”

Introduction to opportunities and pitfalls in functional mass spectrometry based proteomics

“…8 However, the efficiency of statistical comparison between metrics associated to a biological entity is questioned in the literature. [9][10][11][12] Indeed, such studies suffer from the high variance inherently found in biological systems, 9 from the low number of replicates typically analyzed, 10 and from experimental artifacts, errors and missing values. [13][14][15] As a result, the fine nuances of the proteomic variations are often not statistically significant when compared to the global variance of the system.…”