Quick Start
Table of contents
Run metilene3
- Unsupervised mode: with unsupervised mode, only the methylation matrix should be provided.
python /path_to_metilene3/metilene3.py -i methylation.tsv -o /output_dir
Or,
- Supervised mode: with supervised mode, each sample should be assigned to a group, which should be recorded in the
samples_groups.tsv.
python /path_to_metilene3/metilene3.py -i methylation.tsv -g samples_groups.tsv -o /output_dir
More parameters could be added. Please check unsupervised mode and supervised mode for details.
Input file(s)
-
CpG methylation matrix:
methylation.tsv
This matrix should include chromosome, position and CpG methylation levels for each sample. Click to see an example! -
Group information table:
samples_groups.tsv(optional, supervised mode only)
If you already have the group information and wish to compare different groups, you can provide this file to metilene3 with this format.
Output files
-
A HTML report summerizing DMRs and other results:
/output_dir/report.html. -
A tab-separated table containing all DMRs found by metilene3:
/output_dir/DMRs.tsv
(and/output_dir/DMRs-unsupervised.tsvfor DMRs found with unsupervised mode) together with the sample-level DMR methylation rates/output_dir/DMR-met.tsv.
Click here to understand it! -
A tab-separated table containing name and order (ID) for each group:
/output_dir/group-ID.tsv. -
A tab-separated table containing cluster ID for each sample:
/output_dir/clusters.tsv.
(unsupervised mode only) - Figures under the folder
/output_dir/:
(only with unsupervised mode and the parameter visualization-plotis set to True)- DMR tree:
DMTree.[jpg, pdf], showing the hierarchical structure based on the sum of DMR weights (branch lengths) between two branches. You can also find the DMTree in Newick tree format:DMTree.nwk. - PCA 2D-visualization and clustering heatmap based on sample-level DMR methylation ratios:
[PCA, heatmap].[jpg, pdf]. The source data can be found intsvfiles.
- DMR tree:
- GSEA results:
/output_dir/GSEA/(and/output_dir/GSEA-unsupervised/for unsupervised DMRs).
(only if the parameter annotation-annoand the parameter GSEA-gseaare set.)