Methylation Array

This guide provides instructions for analyzing Illumina Infinium Methylation array data. The data set is Infinium Methylation Screening Array Demo Data Set. There are 60 human samples with two different groups and two dose levels in each group.

The input data for methylation array is two .idat files per sample:

<Sample name>_Grn.idat
<Sample name>_Red.idat

Create analysis

After login to Connected Multiomics, create study and add the IDAT files to the study to create 60 samples, add sample meta data.

Click New Analysis button (), specify an analysis name, choose Custom: Illumina Infinium Methylation from Analysis Type drop-down list, and use All Illumina Infinium Methylation samples, click Run Analysis.

A Microarray methylation data node is generated in the Analysis. Click on the data node and choose Generate beta value task.

Generate beta value

Choose Infinium Methylation Screening Array as the Chip name from the drop-down list. Users have the options to exclude probes on XY chromosomes depends on their study goal. To be more conservative, users can also exclude probes if their detect p-values are higher than a cutoff value in certain number of samples. Filter probes will remove noise and speed up the downstream analysis.

After configure the dialog, click Finish to run the task.

Generate PCA

Click on Methylation beta data node, choose PCA in the Exploratory analysis section on the menu.

Keep the default settings and click Finish.

Double click on the PCA data node to open the report in Data viewer.

Detect differential methylation

Single click the Methylation beta data node and perform the Detect differential methylation task under Methylation analysis in the task menu.

This task converts the beta values to m-values and performs ANOVA differential expression analysis on both beta value and m-value matrixes.

Follow along with the task to make one-way or two-way ANOVA comparisons. The configured ANOVA model is performed on both beta value and m-value matrices. Click Finish.

This outputs the Detect differential methylation task report list. Open the Task report from the task menu or double-click the data node.

The outputs of this task include significance as P-value and FDR step up which is from the M-values. The LSMeans of the groups and the Difference are of the Beta-values.

When the difference is greater than or equal to 0.2, it site is labeled as Hypermethylated in Methylation field; if the difference is less than or equal to -0.2, it site is labeled as Hypomethylated in Methylation field. If the difference is between -0.2 to 0.2, it is labeled "?" in Methylation field.

Click the Optional columns button to add more column data including annotation from the Illumina manifest file. For more information on these optional columns, see Infinium Methylation Screening Array Manifest Column Headings on Illumina support site.

Use the left filter panel to filter the results then click Generate filtered node. Detailed information on how to use the filter panel can be found here.

Visualize filtered results with Hierarchical clustering / heatmap

If we use FDR <=0.05 and difference is greater that |0.7| on Cell Line vs Coriell as an example filter criteria to generate a filtered result.

Select the generated Filtered feature list data node and use the Exploratory analysis task menu dropdown to perform the Hierarchical clustering / heatmap task.

Leave the settings in the dialog as default and click Finish.

Double click on the Hierarchical clustering/heatmap to open the report.

Biological interpretation

Select the Filtered feature list data node within the Analyses tab and choose the Gene set enrichment task under Biological interpretation in the task menu. Choose the KEGG database, with other settings as default, click Finish. Double click on the output Pathway enrichment data node to open the report.

When the table is filtered down to less than 100 row, the report can be viewed in Data viewer. The filter can be performed on the report table, or click on the Pathway enrichment data node, perform Filter gene sets in Filtering section. Click on when it is available to open the plots.