This project provides data and code to reproduce the analysis results presented in Jamshidi A., et al., 2022. "Evaluation of Cell-Free DNA Approaches for Multi-Cancer Early Detection"

Follow instructions at https://cran.r-project.org to install R (version 4.1.2 was used for this analysis). The packages needed to reproduce the results have been captured with the package packrat. Pandoc version 2.5 was installed from https://github.com/jgm/pandoc. This implementation was developed and tested on Ubuntu 20.04.3 LTS.

To install the required packages and render the anaysis markdown:

1. Copy the contents of this package (Jamshidi_CCGA1_2022/) to the desired location (/YOUR_LOCATION/Jamshidi_CCGA1_2022).

2. cd into the project directory (/YOUR_LOCATION/Jamshidi_CCGA1_2022), and start R.

3. Install packrat:

install.packages("packrat")

4. Initialize the packrat project:

packrat::init()

Once this has completed, you must quit and restart R for the changes to take effect (https://www.rdocumentation.org/packages/packrat/versions/0.8.0/topics/init).

5. After restarting R, confirm that the packrat library is being used by ensuring there is a startup message like:

Packrat mode on. Using library in directory:
- "/YOUR_LOCATION/Jamshidi_CCGA1_2022/packrat/lib"

Alternatively, from the R console run .libPaths() and ensure that the packrat library is one of the paths (i.e., something like "/YOUR_LOCATION/Jamshidi_CCGA1_2022/packrat/lib/x86_64-apple-darwin17.0/4.1.2" is included).

6. Install all of the versioned dependencies from CRAN and Bioconductor via packrat:

packrat::restore()

Note some packages have dependencies on the user's local setup, and this may necessate troubleshooting if any R packages are not able to installed on the first attempt. For example, RcppArmadillo may require a user install of fortran. These dependencies will differ based on whether run on a Linux machine vs. Mac, which operating system is used, etc. The authors have confirmed that these R packages can be installed and the R markdown can be rendered on both Ubuntu 20.04.3 and macOS Monterey Version 12.5, though different user-specific local installations were required in both cases for all packages to be installed with packrat::restore().

7. Render the R Markdown report containing figures and tables in the publication. Figures (pdf) are saved to /YOUR_LOCATION/Jamshidi_CCGA1_2022/out/, and report .html is saved to /YOUR_LOCATION/Jamshidi_CCGA1_2022/Rmd/make_output.html.

rmarkdown::render("Rmd/make_output.Rmd")

All data required to reproduce the results are provided in the directory data. Each .tsv has a corresponding "dictionary" that contains definitions for each column.

• chip_calls.tsv: Participant-level variant information between cfDNA and WBC gDNA.
• clinical.tsv: Clinical data table.
• lod.tsv: LOD estimates for each Classifier.
• score_and_ctaf.tsv: Classifier score and circulating tumor allele fraction estimates.
• score_cutoffs.tsv: Classifier score cutoffs.
• scores_cnc.tsv: Cancer/Non-cancer Classifier scores for each participant.
• scores_cso.tsv: Cancer Signal Origin Classifier scores for each participant.
• snv_cfdna_assay.tsv: Sample level data for the SNV cfDNA assay.
• snv_wbc_assay.tsv: Sample level data for the SNV WBC assay.
• umap_df_training.tsv: UMAP components of the classifier scores for each participant.
• variant_data_hexbin.tsv: 2D hex-binned data of variant allele frequencies (%) for WBC and cfDNA.
• wgbs_assay.tsv: Sample level data for the WGBS cfDNA assay.
• wgs_cfdna_assay.tsv: Sample level data for the WGS cfDNA assay.
• wgs_wbc_assay.tsv: Sample level data for the WGS WBC assay.