Abstract:Quantitative label-free mass spectrometry (MS) is an increasingly powerful technology for profiling thousands of proteins from complex biological samples. One of the primary goals of analyses performed on such proteomics data is to detect differentially expressed proteins (DEPs) under different experimental conditions. Many statistical methods have been developed and assessed for DEP detection in various proteomics studies. However, it remains a challenge for many proteomics scientists to choose an appropriate statistical procedure. Therefore, in this study, we organized 12 common testing algorithms and 6 P-value combination methods and further provided Cohen's d effect size for every protein and three evaluation criteria to help proteomics scientists investigate their influence on DEP detection in a systematic manner. To promote the widespread use of these methods, we developed a user-friendly web tool, StatsPro, and presented two case studies involving label-free quantitative proteomics data obtained using data-dependent acquisition and data-independent acquisition to illustrate its practicability. This tool is freely available in our GitHub repository (https://github.com/YanglabWCH/StatsPro/). SIGNIFICANCE: One of the primary goals of analyses performed on liquid chromatography-mass spectrometry (LC-MS) based proteomics data is to detect differentially expressed proteins (DEPs) under different experimental conditions. Despite of many research efforts have been proposed to detect DEPs, to date, there is a scarcity of efficient, systematic, and easy-to-handle tools that are tailored for proteomics scientists to choose an appropriate statistical procedure. Herein, we present a new tool, StatsPro, to enable implementation and evaluation of different statistical methods for proteomics scientists. This tool has two significant advances compared to existing software: a) It integrates up to 18 common statistical approaches (12 statistical tests and 6 P-value combination strategies) and performs Cohen's d effect size systematically for users, moreover, it provides a web-based interface and can be quite conveniently operated by users, even those with less profound computational background. b) It supports three performance evaluation criteria (e.g. number of DEPs, correlation coefficient between P-values and effect sizes, Area under the ROC curve) for users to review the final statistical results, which may guide the method selection for DEPs detection.

Statistical Analysis of Multiple Significance Test Methods for Differential Proteomics

StatsPro: Systematic Integration and Evaluation of Statistical Approaches for Detecting Differential Expression in Label-Free Quantitative Proteomics.

Statistical Strategies for Selection of Differentially Expressed Proteins Based on Mass Spectrometry Technology

Comparative evaluation of label-free quantification strategies.

Comparison and Application of Module-Based Statistic Models in Analysis of Protein Profiles

New Statistics for Testing Differential Expression of Pathways from Microarray Data

A two-step strategy for detecting differential gene expression in cDNA microarray data

Simulated linear test applied to quantitative proteomics

Application of multiple statistical tests to enhance mass spectrometry-based biomarker discovery

Using Weighted Permutation Scores to Detect Differential Gene Expression with Microarray Data.

2D DIGE Data Analysis for Differential Protein Spot Identification

Biological Network Module-Based Model for the Analysis of Differential Expression in Shotgun Proteomics

In-depth Method Assessments of Differentially Expressed Protein Detection for Shotgun Proteomics Data with Missing Values

Comparative Assessment and Outlook on Methods for Imputing Proteomics Data

On the Statistical Significance of Protein Complex

Novel Statistical Framework to Identify Differentially Expressed Genes Allowing Transcriptomic Background Differences

A Bayesian Framework for Multivariate Differential Analysis accounting for Missing Data

MAGMa: Your Comprehensive Tool for Differential Expression Analysis in Mass-Spectrometry Proteomic Data.

Significance-Based Essential Protein Discovery

MSstatsPTM: Statistical Relative Quantification of Posttranslational Modifications in Bottom-Up Mass Spectrometry-Based Proteomics

Protein Probabilities in Shotgun Proteomics: Evaluating Different Estimation Methods Using A Semi-Random Sampling Model