Deconvolution

Contents

Deconvolution#

Pattools facilitates multiple deconvolution algorithms tailored to methylation data analysis. The deconvolution algorithms implemented by pattools deconv include:

  • sun [Sun et al., 2015] : based on the methylation densities (i.e., the percentage of CpGs being methylated within a 500-bp unit) of the methylation biomarker

  • moss [Moss et al., 2018] : based on the beta values of the Illumina Infinium Human Methylation 450K or EPIC BeadChip arrays

  • loyfer [Loyfer et al., 2024] : based on the uxm ratios of the DNA regions

usage#

take sun for example

# View the cell components for each method
pattools deconv-helper -m sun 

# run 
pattools deconv -m sun -g hg38 \
    -c /mnt/d/project/wgbs_tools/references/hg38/CpG.bed.gz \
    -p /mnt/d/data/cacLung/raw/pat/GSM5652289_Blood-T-Eff-CD8-Z0000041F.hg38.pat.gz \
    -o /mnt/d/project/pattools/tmp/out.csv \
    --include T-cells B-cells Neutrophils \ 
    --ignore Liver Lung