Samples and Code for Improved deconvolution of circulating tumor DNA from ultra-low-pass whole-genome methylation sequencing using CelFiE-ISH

Dataset Structure epibed_files.zip: Oxford Nanopore sequencing data from 42 samples (20 samples from advanced colon or rectal adenocarcinoma patients and 22 non-cancer control donors) described in manuscript "Improved deconvolution of circulating tumor DNA from ultra-low-pass whole-genome methylation sequencing using CelFiE-ISH" All data are aligned to UCSC analysisSet hg38 (https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/analysisSet/), and use 0-based coordinates. To provide read-level DNA methylation information, Biscuit was used to generate epiBED files (see Methods). epiBED files represent each read on a separate line and encode all DNA methylation calls along the read, enabling integrated methylation analyses. These files are compatible with the CelFiE-ISH framework. Creation of Biscuit EpiBED files: Hydroxymethylated bases in BAM files were converted to methylated bases using modkit `modkit modbam adjust-mods` command and the `--convert h m` option. Hard clipping was removed from the start and end of CIGAR strings, and empty methylation tags were added to ensure compatibility with Biscuit. Epireads were generated using Biscuit v1.7.1 with the command “biscuit epiread -M -b 0 -m 0 -a 0 -5 0 -3 0 -y 0.9 -L 2300000 -E hg38.analysisSet.fa”, using the same UCSC reference genome as for alignment. Variant run-length–encoded strings were masked. run_celfie_ish.zip Contains the input files used for the CelFiE-ISH deconvolution framework as described in the paper. code.zip: Includes code and analysis scripts used to generate the results and figures presented in the manuscript.