This is a standalone program for deconvolution of array methylome[1]. It uses an input reference atlas file to deconvolve a given sample, or multiple samples. Outputs a csv file, and plots a stacked bars figure.
A reference atlas file.
The reference atlas used on the paper is supplied in this repository - reference_atlas.csv. The full reference atlas, with ~390K sites (before the feature selection process) is also supplied - full_reference_atlas.csv.gz. For better deconvolurions results, it's recommended not to use the full atlas, but the smaller one.
A file containing one or more samples, with similar requirements as the atlas file (csv, header, index column). The CpG (Illumina ID) column may contain different CpG sites than the ones in the atlas files, as long as they share some sites.
An example dummy file is supplied, examples.csv.
usage: deconvolve.py [-h] [--atlas_path ATLAS_PATH] [--slim] [--plot]
[--out_dir OUT_DIR]
samples_path
positional arguments:
samples_path Path to samples csv file. It must have a header line.
The first column must be Illumina IDs (e.g cg00000029)
optional arguments:
-h, --help show this help message and exit
--atlas_path ATLAS_PATH, -a ATLAS_PATH
Path to Atlas csv file. The first column must be
Illumina IDs (e.g cg00000029)
--slim Write the results table *without indexes and header
line*
--plot Display the stacked bars figure
--out_dir OUT_DIR, -o OUT_DIR
Output directory
deconvolve.py -a reference_atlas.csv examples.csv
will deconvolve all samples given as columns in examples.csv, dump the resulting coefficients to a csv file named examples_deconv_output.csv, plot them, and dump the figure to examples_deconv_plot.png.
We recommend pre-processing the IDAT files using the script and reference sample listed here.
[1]: Moss, J.; Magenheim, J.; Neiman, D.; Zemmour, H.; Loyfer, N.; Korach, A.; Samet, Y.; Maoz, M.; Druid, H.; Arner, P.; et al. Comprehensive human cell-type methylation atlas reveals origins of circulating cell-free DNA in health and disease. Nat. Commun. 2018, 9, 5068.