Differential Expression Analysis

Table of contents

Introduction

In -omic studies, volcano plot is often used to show differential expressions of features (proteins or genes) between groups. Differential expressions can be concluded through two parameters: fold-change between groups and p value of hypothesis test.

In a volcano plot, the x-axis is usually the Log2 fold-change of each feature. A feature with higher fold change is farther off the center along the x-axis. The y-axis shows the -Log10(p-value). For high dimensional data, it is recommended that the p value is adjusted (i.e. p-adjust). A feature that changes with a greater statistical significance has a larger value along the y-axis. Many features have both high number of fold change and a great significance of change, and thus make the scatter plot look like a volcanic eruption.

Types of volcano plots

The most commonly seen volcano plots use right-angle threshold, which is determined by a fixed number of fold-change and an adjusted p value. In our Differential Protein Analyzer , we also introduced a curve threshold volcano plot (inspired by Keilhauer et al., Mol Cell Proteomics. 2015 Jan; 14(1): 120-135). The curve is determined by the following function:

y > curvature / |x-“Log2FoldChangeCutOff”| + “-Log10pValueCutOff”

In this app, parametric test is performed using function t.test(), Non-parametric test is performed using function wilcox.test(), p values are adjusted using function p.adjust(). Result is visualized using R packages “ggplot2” and “ggplotly”.

Tutorial

Step 1. Prepare your data tables

Two tables are required for the differential protein analysis: data table, and meta table. The data table is a two-dimensional intensity table with rows of features and columns of samples. The meta table includes sample names matching the data table, and group information of each sample.

Step 2. Upload data tables

Go to https://shiny.imetalab.ca/Volcano_plot/ , upload data matrix and meta table separately. Select the correct format of your data table in the dialogue. If data table are uploaded correctly, there will be a prompt to double-check data table.

Step 3. Pre-process data

If your data needs to be processed before volcano plot, turn on the “Process data before analysis” button shown below. There, you can perform filtering, imputation, transformation and normalization on your data table.

If your data doesn’t need processing, you can directly click the “Go to differential expression analysis button”.

Step 4. Draw volcano plot

After clicking the go to analysis button, a volcano plot with our default settings (recommended) will appear automatically.

You can change the following statistical settings according to your requirements.

In the advanced tab, you can also change the figure appearance.

The result table is downloadable for your further analyses.

Back to the list of iMetaShiny Apps.