Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 May 6;5(1):vbaf108.
doi: 10.1093/bioadv/vbaf108. eCollection 2025.

cfTools: an R/Bioconductor package for deconvolving cell-free DNA via methylation analysis

Ran Hu  1   2   3 , Shuo Li  1 , Mary L Stackpole  1 , Qingjiao Li  1 , Xianghong Jasmine Zhou  1   3 , Wenyuan Li  1
Affiliations

cfTools: an R/Bioconductor package for deconvolving cell-free DNA via methylation analysis

Ran Hu et al. Bioinform Adv. .

Abstract

Motivation: Cell-free DNA (cfDNA) released by dying cells from damaged or diseased tissues can lead to elevated tissue-specific DNA, which is traceable and quantifiable through unique DNA methylation patterns. Therefore, tracing cfDNA origins by analyzing its methylation profiles holds great potential for detecting and monitoring a range of diseases, including cancers. However, deconvolving tissue-specific cfDNA remains challenging for broader applications and research due to the scarcity of specialized, user-friendly bioinformatics tools.

Results: To address this, we developed cfTools, an R package that streamlines cfDNA tissue-of-origin analysis for disease detection and monitoring. Integrating advanced cfDNA tissue deconvolution algorithms with R/Bioconductor compatibility, cfTools offers data preparation and analysis functions with flexible parameters for user-friendliness. By identifying abnormal cfDNA compositions, cfTools can infer the presence of underlying pathological conditions, including but not limited to cancer. It simplifies bioinformatics tasks and enables users without advanced expertise to easily derive biologically interpretable insights from standard preprocessed sequencing data, thus increasing its accessibility and broadening its application in cfDNA-based disease studies.

Availability and implementation: cfTools and its supplementary package cfToolsData are freely available at Bioconductor: https://bioconductor.org/packages/release/bioc/html/cfTools.html and https://bioconductor.org/packages/release/data/experiment/html/cfToolsData.html. The development version of cfTools is maintained on GitHub: https://github.com/jasminezhoulab/cfTools.

PubMed Disclaimer

Conflict of interest statement

X.J.Z. and W.L. are co-founders of EarlyDiagnostics, Inc. X.J.Z. serves on the Board of Directors and has an executive leadership position at EarlyDiagnostics. W.L. serves as the Board of Directors at EarlyDiagnostics. W.L. and X.J.Z. are stockholders of EarlyDiagnostics. M.L.S. is an employee of EarlyDiagnostics and has stock options with EarlyDiagnostics. W.L. and S.L. are consultants to EarlyDiagnostics and S.L. has stock options with EarlyDiagnostics. The other authors have no competing interests to declare.

Figures

Figure 1.
Figure 1.
The cfTools workflow and key data files. (a) The main workflow of cfTools encompassing data processing and downstream analysis. The data preparation functions for generating fragment-level methylation states take standard output files from Bismark as inputs. To generate marker parameter files, methylation levels of markers across a list of samples are required. CancerDetector and cfDeconvolve statistically model the methylation patterns of cancer or tissue marker to deconvolve tumor-derived or tissue-specific fragments, thereby estimating tumor burden or tissue fraction of user-defined tissue types. cfSort infers the composition of 29 predefined tissues from the fragment-level methylation profile of a cfDNA sample. (b) The output file of the GenerateFragMeth function contains both fragment information and the methylation states of all CpGs on the fragment. The fragment-level methylation states are in the last column “methState”. (c) The output file of the GenerateMarkerParam function contains the paired shape parameters of beta distributions for each marker across different tissues. The parameters of a Beta(α, β) distribution are formatted as “α:β”.
Figure 2.
Figure 2.
Box plots of diseased tissue-derived cfDNA fraction from cfDeconvolve and cfSort, and tumor burden from CancerDetector in patients and healthy people. (a–c) The estimated lung-derived cfDNA fractions and tumor burden in lung cancer patients versus normal individuals. (d–f) The estimated liver-derived cfDNA fractions and tumor burden in liver cancer patients versus normal individuals. (g and h) The estimated liver-derived cfDNA fractions in patients with cirrhosis versus normal individuals. The distinction between patients and normal individuals was assessed through the Wilcoxon rank-sum tests. The statistical significance of the tests was denoted by the asterisks: “****” means P value < .0001, “***” means P value < .001.
See this image and copyright information in PMC

References

    1. Barefoot ME, Loyfer N, Kiliti AJ et al. Detection of cell types contributing to cancer from circulating, cell-free methylated DNA. Front Genet 2021;12:671057. - PMC - PubMed
    1. Caggiano C, Celona B, Garton F et al. Comprehensive cell type decomposition of circulating cell-free DNA with CelFiE. Nat Commun 2021;12:2717. - PMC - PubMed
    1. Del Vecchio G, Li Q, Li W et al. Cell-free DNA methylation and transcriptomic signature prediction of pregnancies with adverse outcomes. Epigenetics 2021;16:642–61. - PMC - PubMed
    1. Feng H, Jin P, Wu H et al. Disease prediction by cell-free DNA methylation. Brief Bioinform 2019;20:585–97. - PMC - PubMed
    1. Huang J, Wang L. Cell-free DNA methylation profiling analysis—technologies and bioinformatics. Cancers (Basel) 2019;11:1741. - PMC - PubMed