▸ Help


  ▸ eFORGE is the epigenetic equivalent of FORGE, using EWAS rather than GWAS data.

▸ eFORGE identifies tissue or cell type-specific signal by analysing a minimum set of 5 differentially methylated positions (DMPs) for overlap with DNase I hypersensitive sites (DHSs) compared to matched background DMPs and provides both graphical and tabulated outputs.

▸ eFORGE can now analyse Illumina 850k EPIC array data, as well as Illumina 450k array data.

Input data

  ▸ Source

You can upload a file, provide a URL for an existing file or simply copy " paste your data on the text box provided.

▸ Input file format

If f is specified, specify the file format as follow:

Probe list: list of dmps as probeids each on a separate line. Optionally can add other fields after the probeid which are ignored, unless the pvalue filter is specified, in which case eForge assumes that the second field is the minus log10 pvalue

BED file: File given is a bed file of locations (chr beg end). bed format should be 0 based and the chromosome should be given as chrN. However we will also accept chomosomes as just N (ensembl) and 1-based format where beg and end are the same*.

▸ Platform

Select which Illumina methylation array has been used to generate these data (850k or 450k).

▸ Maximum number of probes

The web version is limited to 1000 probes. While it is possible to input a larger number of probes with the standalone version, you must consider how the background selection occurs.


  ▸ Name for this data

Supply a label that you want to use for the plotting titles, and filenames.

▸ Analysis data from

Dataset to analyse. Either ENCODE data ('encode') or Roadmap Epigenome data ('erc'). erc by default.

▸ Depletion

Analyse for DHS depletion pattern instead of the default DHS enrichment analysis. Use when dealing with datasets suspected not to overlap with DHS. Specifying depletion will be indicated on the label (the text "Depletion Analysis" will be added to the file label).

▸ Proximity

Apply filter for DMPs in proximity (within 1 kb of another test DMP). With proximity filter specified, eFORGE will report DMPs removed due to proximity with another DMP in the list and will randomly pick one of the probes among the set of probes that are in proximity (within 1 kb of each other).

▸ Background repetitions

The number of background matching sets to pick and analyse. (1-1000)

▸ Significance threshold

Alter the default binomial p value thresholds. (0 < Strict < Marginal < 1)


  ▸ Raw data

A tab-separated values (tsv) file with columns for the Zscore, Pvalue, Cell, Tissue, File, Probe, Number and Accession.

▸ Static chart

A PDF chart of the data.

▸ Interactive chart

A dimple (http://dimplejs.org) d3 interactive graphic using rCharts.

▸ Interactive table

A table using the Datatables (https://datatables.net) plug-in for the jQuery Javascript library, again accessed through rChartsA dimple (http://dimplejs.org) d3 interactive graphic using rCharts.

▸ R code

The source files for generating the charts and the table. You can download the raw data and use R (http://www.r-project.org) with these file to re-generate or modify the output.