Sort by

Top genes

Download results

Complete table

Filtered table



Selected datasets

Automatic datasets selection



Show covariates

Sort datasets by covariates

FDR filter

Height (in pixels)

Download plot

Download data



Top genes


ADEx (Autoimmune Diseases Explorer) is a tool for easy exploration and analysis of expression and methylation data from a large cohort of systemic autoimmune diseases (SADs), including data from studies with a case-control design with samples affected by systemic lupus erythematosus (SLE), rheumatoid arthritis (RA) and Sjögren’s syndrome (SjS). Data can be explored at three different levels:

  1. Gene level: Expression and methylation data of one or few genes can be explored in one dataset.
  2. Dataset level: Differential expression analysis, pathway and network-analysis can be carried out in individual datasets.
  3. Meta-analysis: Data from different datasets are integrated and analyzed to define common differentially expressed genes.


All data included in ADEx has been obtained from the public repository GEO [1]. We manually curated each dataset to obtain proper expression, methylation and clinical information. If a dataset GEO contained samples from different origins/diseases, wesplit those datasets with samples from more than one disease (e.g. GSE4588 was split into GSE4588_SLE and GSE4588_RA). In addition, datasets with more than one platform were split too (e.g. GSE11907 was split into GSE11907_1 and GSE11907_2). Table 1 contains information for all the included datasets.

Table 1. Datasets available in ADEx.


You can change between different levels of data analysis by clicking on the different tabs at the top of the page (Figure 1). Inside each section there are different tabs to select a specific kind of analysis into the chosen category.

Figure 1: ADEx tabs

In the following sections each tab and subtab are explained.

3.1. Data overview

In this section you can get information for all datasets individually or grouped by disease. To select the datasets, use the left panel.

3.1.1. Table

Here there is an interactive table with the available information for each sample from the selected dataset(s). Samples can be ordered by any column and there is a searcher on the top-right of the table.

3.1.2. Plot

In this subtab a pie chart is generated showing the distribution of samples in the selected dataset(s) (Figure 2). Samples are grouped by the variable selected in the right panel.

Figure 2: Pie chart

3.2. Gene-centered analysis

In this section, expression and methylation data from a chosen gene can be represented in different plots. Datasets and genes can be chosen in the left panel.

3.2.1. Expression

A boxplot with the expression values in each condition. On the top of the plot there is the P-value of Wilcoxon's test between conditions, showing the evidence of differential expression.

3.2.2. Methylation

At the top of this page, you can select the region of the gene to plot methylation values (Figure 3). You can automatically select the promoter, the gene body or the promoter and gene body clicking on the corresponding buttons.

Figure 3: Region selector for methylation data

A boxplot and a lineplot are generated in this section showing the methylation values for each CpG into the selected region. These plots can be scaled by the CpGs' genomic positions checking the corresponding option on the right panel.

3.2.3. Expression + Methylation

Here you can integrate expression and methylation data for the datasets containing these two kinds of data. A correlation plot is generated. Each point of the plot represents the expression of a sample on the y-axis and the mean methylation value of all the CpGs into the selected region of a sample on the x-axis. Correlation method and regression line can be controlled in the right panel.

3.2.4. Expression correlation

In this analysis, correlation between all the selected genes are plotted. Genes are clustered in order to get groups of correlated genes (Figure 4).

At the top of this page, you can select the region of the gene to plot methylation values (Figure 3). You can automatically select the promoter, the gene body or the promoter and gene body clicking on the corresponding buttons.

Figure 4: Correlation plot

3.3. Dataset-centered analysis

In this section, different analysis can be performed for a chosen dataset.

3.3.1. Differential Expression

Differential expression analysis between diseased and healthy samples is performed using limma package [2]. The results of such analysis are represented in a heatmap. Heatmap shows the top genes ordered by P-value or by Fold-Change. Both ordering metric and top genes can be specified on the right panel. Colors in the heatmap are assigned based on the expression values scaled in each gene.

3.3.2. Pathway analysis

Lists of significant differentially expressed genes (FDR < 0.05) from the previous limma analysis are used to perform an enrichment analysis in KEGG pathways. The results of this pathway analysis are showed in an interactive table. Plots on the last column can be clicked to open a new window with a full-size plot. In these plots, significant genes are colored by their Fold-Change in the comparison.

3.3.3. Signal transduction

Canonical Circuit Activity Analysis [3] is performed for each dataset. In this kind of analysis, pathways are decomposed into circuits. The activity of different circuits determines the activation or repression of pathway functions. A table with a summary of the results is shown in this tab. In addition, clicking on “Open interactive report” button allows to open a new window with an interactive report, were different pathways and circuits can be dynamically visualized (Figure 5).

Figure 5: Signal transduction report.

3.4. Dataset Meta-analysis

In this tab, different kinds of meta-analysis can be performed with all the desired datasets.

3.4.1. Datasets selection

This is the subtab where the datasets are chosen to perform the analysis of all the other subtabs. You can select as many samples as you want clicking on their rows. In addition, buttons in the left panel can be used to automatically select/unselect all datasets or the datasets with samples of a disease.

3.4.2. Gene expression meta-analysis

Here, you first have to select a comparison to perform the meta-analysis and, then, press “Launch Meta-Analysis”. Take into account that this process can be slow. Meta-analysis is performed aggregating the corrected P-values from the individual differential expression analysis using Fisher method [4]. The results is showed in a heatmap similar to the one generated in the individual analysis.

3.4.3. Differential expression meta-analysis

Here you can see in an interactive plot the P-value, FDR or Fold-Change of the desired genes in the selected datasets. The to be showed is selected in the right panel.

3.4.4. Pathway meta-analysis

This is similar to the previous pathway analysis, but considering as significant differentially expressed genes those with a FDR < 0.05 in the gene expression meta-analysis. Take into account that, given that this analysis depends on the gene expression meta-analysis, the later must be done before performing the pathway meta-analysis.


  1. Edgar R, Domrachev M, Lash AE (2002). Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Research, 30(1):207-10
  2. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK (2015). limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Research, 43(7): e47
  3. Hidalgo MR, Çubuk C, Amadoz A, Salavert F, Carbonell-Caballero J, Dopazo J (2017). High throughput estimation of functional cell activities reveals disease mechanisms and predicts relevant clinical outcomes. Oncotarget, 3(8), 5160–5178
  4. Sutton AJ, Abrams, KR, Jones DR, Sheldon TA, Song, F (2000). Methods for meta-analysis in medical research. Wiley, Chichester


If you have any doubt, question or suggestion, you can write us to .