. 2022 Sep 29;13(1):5566.

doi: 10.1038/s41467-022-32995-6.

Cost-effective methylome sequencing of cell-free DNA for accurately detecting and locating cancer

Mary L Stackpole^#^{1

2}, Weihua Zeng^#¹, Shuo Li^#^{1

2}, Chun-Chi Liu², Yonggang Zhou¹, Shanshan He¹, Angela Yeh¹, Ziye Wang¹, Fengzhu Sun³, Qingjiao Li⁴, Zuyang Yuan¹, Asli Yildirim⁵, Pin-Jung Chen⁶, Paul Winograd⁶, Benjamin Tran⁶, Yi-Te Lee⁷, Paul Shize Li⁸, Zorawar Noor⁹, Megumi Yokomizo¹⁰, Preeti Ahuja^{10

11}, Yazhen Zhu^{7

11}, Hsian-Rong Tseng^{7

11}, James S Tomlinson^{6

9

11

12}, Edward Garon^{9

11}, Samuel French^{1

11}, Clara E Magyar^{1

11}, Sarah Dry^{1

11}, Clara Lajonchere^{11

13}, Daniel Geschwind¹³, Gina Choi⁹, Sammy Saab⁹, Frank Alber^{5

14}, Wing Hung Wong^{15

16}, Steven M Dubinett^{1

7

9

11

12}, Denise R Aberle^{10

11}, Vatche Agopian^{17

18}, Steven-Huy B Han¹⁹, Xiaohui Ni²⁰, Wenyuan Li²¹, Xianghong Jasmine Zhou^{22

23

24}

Affiliations

¹ Department of Pathology and Laboratory Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
² EarlyDiagnostics, Inc., 570 Westwood Plaza, Los Angeles, CA, 90095, USA.
³ Department of Quantitative and Computational Biology, University of Southern California, Los Angeles, CA, 90089, USA.
⁴ The Eighth Affiliated Hospital, Sun Yat-Sen University, Shenzhen, China.
⁵ Department of Microbiology, Immunology and Molecular Genetics, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
⁶ Department of Surgery, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
⁷ Department of Molecular and Medical Pharmacology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
⁸ Westlake High School, 100N Lakeview Cyn Road, Westlake Village, CA, 91362, USA.
⁹ Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹⁰ Department of Radiological Sciences, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹¹ Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹² VA Greater Los Angeles Health Care System, Los Angeles, CA, 90073, USA.
¹³ Institute for Precision Health, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹⁴ Institute for Quantitative and Computational Biosciences, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹⁵ Department of Statistics, Stanford University, Stanford, CA, 94305, USA.
¹⁶ Department of Biomedical Data Science, Stanford University, Stanford, CA, 94305, USA.
¹⁷ Department of Surgery, University of California at Los Angeles, Los Angeles, CA, 90095, USA. vagopian@mednet.ucla.edu.
¹⁸ Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, CA, 90095, USA. vagopian@mednet.ucla.edu.
¹⁹ Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA. steven.han@ucla.edu.
²⁰ EarlyDiagnostics, Inc., 570 Westwood Plaza, Los Angeles, CA, 90095, USA. xiaohui.ni@earlydx.com.
²¹ Department of Pathology and Laboratory Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA. wenyuanli@mednet.ucla.edu.
²² Department of Pathology and Laboratory Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA. xjzhou@mednet.ucla.edu.
²³ Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, CA, 90095, USA. xjzhou@mednet.ucla.edu.
²⁴ Institute for Quantitative and Computational Biosciences, University of California at Los Angeles, Los Angeles, CA, 90095, USA. xjzhou@mednet.ucla.edu.

^# Contributed equally.

PMID: 36175411
PMCID: PMC9522828
DOI: 10.1038/s41467-022-32995-6

Cost-effective methylome sequencing of cell-free DNA for accurately detecting and locating cancer

Mary L Stackpole et al. Nat Commun. 2022.

. 2022 Sep 29;13(1):5566.

doi: 10.1038/s41467-022-32995-6.

Authors

Affiliations

¹ Department of Pathology and Laboratory Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
² EarlyDiagnostics, Inc., 570 Westwood Plaza, Los Angeles, CA, 90095, USA.
³ Department of Quantitative and Computational Biology, University of Southern California, Los Angeles, CA, 90089, USA.
⁴ The Eighth Affiliated Hospital, Sun Yat-Sen University, Shenzhen, China.
⁵ Department of Microbiology, Immunology and Molecular Genetics, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
⁶ Department of Surgery, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
⁷ Department of Molecular and Medical Pharmacology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
⁸ Westlake High School, 100N Lakeview Cyn Road, Westlake Village, CA, 91362, USA.
⁹ Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹⁰ Department of Radiological Sciences, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹¹ Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹² VA Greater Los Angeles Health Care System, Los Angeles, CA, 90073, USA.
¹³ Institute for Precision Health, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹⁴ Institute for Quantitative and Computational Biosciences, University of California at Los Angeles, Los Angeles, CA, 90095, USA.
¹⁵ Department of Statistics, Stanford University, Stanford, CA, 94305, USA.
¹⁶ Department of Biomedical Data Science, Stanford University, Stanford, CA, 94305, USA.
¹⁷ Department of Surgery, University of California at Los Angeles, Los Angeles, CA, 90095, USA. vagopian@mednet.ucla.edu.
¹⁸ Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, CA, 90095, USA. vagopian@mednet.ucla.edu.
¹⁹ Department of Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA. steven.han@ucla.edu.
²⁰ EarlyDiagnostics, Inc., 570 Westwood Plaza, Los Angeles, CA, 90095, USA. xiaohui.ni@earlydx.com.
²¹ Department of Pathology and Laboratory Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA. wenyuanli@mednet.ucla.edu.
²² Department of Pathology and Laboratory Medicine, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, 90095, USA. xjzhou@mednet.ucla.edu.
²³ Jonsson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, CA, 90095, USA. xjzhou@mednet.ucla.edu.
²⁴ Institute for Quantitative and Computational Biosciences, University of California at Los Angeles, Los Angeles, CA, 90095, USA. xjzhou@mednet.ucla.edu.

^# Contributed equally.

PMID: 36175411
PMCID: PMC9522828
DOI: 10.1038/s41467-022-32995-6

Erratum in

Author Correction: Cost-effective methylome sequencing of cell-free DNA for accurately detecting and locating cancer.
Stackpole ML, Zeng W, Li S, Liu CC, Zhou Y, He S, Yeh A, Wang Z, Sun F, Li Q, Yuan Z, Yildirim A, Chen PJ, Winograd P, Tran B, Lee YT, Li PS, Noor Z, Yokomizo M, Ahuja P, Zhu Y, Tseng HR, Tomlinson JS, Garon E, French S, Magyar CE, Dry S, Lajonchere C, Geschwind D, Choi G, Saab S, Alber F, Wong WH, Dubinett SM, Aberle DR, Agopian V, Han SB, Ni X, Li W, Zhou XJ. Stackpole ML, et al. Nat Commun. 2024 May 1;15(1):3693. doi: 10.1038/s41467-024-48018-5. Nat Commun. 2024. PMID: 38693151 Free PMC article. No abstract available.

Abstract

Early cancer detection by cell-free DNA faces multiple challenges: low fraction of tumor cell-free DNA, molecular heterogeneity of cancer, and sample sizes that are not sufficient to reflect diverse patient populations. Here, we develop a cancer detection approach to address these challenges. It consists of an assay, cfMethyl-Seq, for cost-effective sequencing of the cell-free DNA methylome (with > 12-fold enrichment over whole genome bisulfite sequencing in CpG islands), and a computational method to extract methylation information and diagnose patients. Applying our approach to 408 colon, liver, lung, and stomach cancer patients and controls, at 97.9% specificity we achieve 80.7% and 74.5% sensitivity in detecting all-stage and early-stage cancer, and 89.1% and 85.0% accuracy for locating tissue-of-origin of all-stage and early-stage cancer, respectively. Our approach cost-effectively retains methylome profiles of cancer abnormalities, allowing us to learn new features and expand to other cancer types as training cohorts grow.

PubMed Disclaimer

Conflict of interest statement

X.J.Z., W.L., and W.H.W. are co-founders of EarlyDiagnostics, Inc. X.J.Z and W.H.W are board members for EarlyDiagnostics, Inc. X.J.Z. has an executive leadership position at EarlyDiagnostics, Inc. M.L.S, X.N., and C-C.L. are employees of EarlyDiagnostics,Inc and S.M.D. was a scientific advisor to EarlyDiagnostics, Inc. X.J.Z., W.L., W.H.W., and F.A. are stockholders of EarlyDiagnostics, Inc. M.L.S, W.Z., S.L., C.-C.L., Y.Zhou, Q.L., X.N. have stock options with EarlyDiagnostics, Inc. W.L., W.Z., and Y.Zhou are consultants for EarlyDiagnostics, Inc. X.J.Z., M.L.S, Y.Zhou, X.N., and W.Z. are inventors on a patent application submitted by the Regents of the University of California and licensed to EarlyDiagnostics, Inc. (Patent No. US20210404007A1). P.S.L. performed summer internships in EarlyDiagnostics, Inc. in 2021 and 2022. The other authors have no competing interests to declare.

Figures

**Fig. 1. cfMethyl-Seq assay.**
a Diagram of the cfMethyl-Seq protocol. b Typical TBE-UREA PAGE image of cfMethyl-Seq libraries made from cfDNA, compared with conventional RRBS with cfDNA or intact genomic DNA as input material. The non-specific ligation product from cfDNA fragments with the conventional RRBS protocol is indicated by an arrow. This technical validation experiment was repeated independently twice and showed similar results. For cfMethyl-Seq assays generating data for analysis, each sample was constructed into library without replicate. c The percentage of reads with MspI sites on both ends, on only one end, and on neither end from WGBS assay, our cfMethyl-Seq assay, and RRBS assay on cfDNA. Source data are provided as a Source Data file. d The percentage of mapped fragments that fall in CpG islands, CpG island shores, CpG island shelves, and open sea regions is shown for cfMethyl-Seq libraries, RRBS libraries, and WGBS libraries on cfDNA. Source data are provided as a Source Data file. e Methylation concordance between a genomic DNA sample sequenced with RRBS, and sheared and sequenced with cfMethyl-Seq, increases with depth of coverage. Pearson correlation (y-axis) of the methylation rate (beta value) in the two datasets was calculated on the CpG sites that are covered by both datasets at minimum depth of coverage specified on the x-axis. Source data are provided as a Source Data file. Abbreviations: RRBS Reduced representation bisulfite sequencing.

**Fig. 2. Study design and overview of the computational method.**
a Overview of the sample usage for marker discovery, model training, and validation. All tissue samples are used for marker discovery, and all plasma samples are randomly split into three sets, used for marker discovery, training, and validating the predictive model. The plasma sample split is repeated 10 times and the prediction performance is averaged over the 10 runs. b Details of sample usage for marker discovery. Different types of methylation markers were discovered by using different samples. Note that 30 reference noncancer plasma samples (in blue boxes) correspond to “marker filtration” in a. Abbreviations: TOO tissue of origin.

**Fig. 3. Performance of the stacked ensemble model for cancer detection.**
a ROC curve of the stacked ensemble method for detecting all four cancer types. Source data are provided as a source data file. b Sensitivity breakdown in each cancer stage and cancer type. Sensitivity is shown at 1 false positive (97.9% specificity). The average number of test cancer patients in each cancer type and stage over 10 runs is indicated in the label of the horizontal axis. Sensitivity is not computed if the average number of cancer patients in a cancer stage/type over 10 runs is <4. The points and error bars represent the average sensitivity over 10 runs and 95% confidence intervals. Source data are provided as a Source Data file. c Performance (AUROC) of using all marker types and each individual marker type ( $n :$ 102 samples). The points and error bars represent the average AUROC over 10 runs and 95% confidence intervals. Source data are provided as a Source Data file.

**Fig. 4. Performance of the stacked ensemble model for cancer Tissue-Of-Origin prediction.**
a Confusion matrix for all-stage cancer samples. Source data are provided as a Source Data file. b Confusion matrix for early-stage (i.e., stage I/II) cancer samples. Source data are provided as a Source Data file. c The accuracy of using all marker types and each individual marker type ( $n :$ 35 samples in the test set of each run). The points and error bars represent the average accuracy over 10 runs and 95% confidence intervals. Source data are provided as a Source Data file.

**Fig. 5. Impact of the number of markers and the training sample size on the cancer detection performance.**
a Performance of using the union of top M cancer-specific markers of four cancer types. Source data are provided as a Source Data file. b Performance of using the union of top M tissue-specific markers of each tissue pair. Source data are provided as a Source Data file. c Performance of the ensemble model for cancer detection increases with increasing training sample size (using 30% to 100% of the training samples). Source data are provided as a Source Data file. In all figures, the points and error bars represent the average AUROC over 10 runs and 95% confidence intervals ( $n :$ 102 test samples per run).

See this image and copyright information in PMC

References

1. Cohen JD, et al. Detection and localization of surgically resectable cancers with a multi-analyte blood test. Science. 2018;359:926–930. doi: 10.1126/science.aar3247. - DOI - PMC - PubMed
1. Liu MC, et al. Sensitive and specific multi-cancer detection and localization using methylation signatures in cell-free DNA. Ann. Oncol. 2020;31:745–759. doi: 10.1016/j.annonc.2020.02.011. - DOI - PMC - PubMed
1. Cristiano S, et al. Genome-wide cell-free DNA fragmentation in patients with cancer. Nature. 2019;570:385–389. doi: 10.1038/s41586-019-1272-6. - DOI - PMC - PubMed
1. Guo S, et al. Identification of methylation haplotype blocks aids in deconvolution of heterogeneous tissue samples and tumor tissue-of-origin mapping from plasma DNA. Nat. Genet. 2017;49:635–642. doi: 10.1038/ng.3805. - DOI - PMC - PubMed
1. Xu, R.-H. et al. Circulating tumour DNA methylation markers for diagnosis and prognosis of hepatocellular carcinoma. Nat. Mater.10.1038/nmat4997 (2017). - PubMed

Publication types

Actions
Actions
Actions

MeSH terms

Actions
Actions
Actions
Actions
Actions
Actions
Actions

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

Cost-effective methylome sequencing of cell-free DNA for accurately detecting and locating cancer

Affiliations

Cost-effective methylome sequencing of cell-free DNA for accurately detecting and locating cancer

Authors

Affiliations

Erratum in

Abstract

Conflict of interest statement

Figures

References

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources

Full Text Sources

Other Literature Sources

Medical