Skip to Main Content

An official website of the United States government

Screen early-stage cancers by single-cell multi-omics and cell-free DNA fragmentation patterns

Principal Investigator
Name
Yaping Liu
Degrees
Ph.D.
Institution
Northwestern University Feinberg School of Medicine
Position Title
Assistant Professor
Email
About this CDAS Project
Study
PLCO (Learn more about this study)
Project ID
2024-0111
Initial CDAS Request Approval
Aug 15, 2024
Title
Screen early-stage cancers by single-cell multi-omics and cell-free DNA fragmentation patterns
Summary
Screening early-stage cancer provides the opportunity to reduce morbidity and mortality in patients. The current gold standard to diagnose cancer heavily relies on needle biopsy, which has limited power for early-stage cancer due to the tiny volume of solid tumors in the very early stage. Recent advances in liquid biopsy, such as circulating cell-free DNA (cfDNA), provided a promising non-invasive approach to diagnose early-stage cancer. The cfDNA fragmentation and their derived patterns, such as nucleosome protection, offer extensive interpretable signals from both tumor and immune cells and thus showed significant increases of sensitivity for cancer early detection over other approaches. Most state-of-the-art studies of cancer early detection by cfDNA focused on the samples from cancer patients at the time of diagnosis rather than those before diagnosis in screening. Patients may already show symptoms months or even years before getting referred for a conventional cancer diagnosis. Here, we propose to apply a novel single-cell multi-omics approach we recently developed at white blood cells to discover the potential regions of interest in the genome and relevant immune cell types that are aberrant at the time before conventional cancer diagnosis. Further, we will screen multiple early-stage cancers in the average-risk population by cfDNA fragmentation patterns from these regions at the time before conventional diagnosis.
Aims

Routine screening in the average-risk population provides the opportunity to detect cancer early and significantly reduce morbidity and mortality in cancer. Recent advances in circulating cell-free DNA (cfDNA) suggested a promising non-invasive approach for cancer diagnosis by using tumor-specific genetic and epigenetic alterations. However, in very early-stage cancers, most cfDNA fragments are released from peripheral immune cells but not tumors. Epigenetic marks play critical roles in cancer initiation and may already show aberrations in peripheral immune cells even before cancer is diagnosed. However, how the epigenetic marks were changed in white blood cells in the very early stage and what the relevant cell types are have not been well studied, especially in granulocytes which released most of the cfDNA fragments. We recently developed a novel single-cell multi-omics approach to jointly profile DNA methylation, chromatin accessibility, and 3D genome within the same nuclei at flash-frozen cells, which provides an opportunity to understand the etiology aberrations in peripheral immune cells at the single-cell level. Moreover, we and others also discovered the correlation between cfDNA fragmentation and epigenetic marks within the cells that contribute to cfDNA. Therefore, we hypothesize that the epigenetic aberrations we discovered in peripheral immune cells from early-stage cancer patients can be utilized to boost the power of cancer early detection by cell-free DNA fragmentation pattern, even at the stage before traditional cancer diagnosis.
Aim 1. Discover the epigenetic aberrations from peripheral immune cells in early-stage cancers by single-cell multi-omics
Traditional single-cell epigenetic approaches, such as single-cell ATAC-seq, can not be applied to granulocytes, which release most of the cfDNA fragments. Our recently developed single-cell NOMe-HiC (unpublished, derived from Fu et al. 2023 Genome Biol) can be applied to flash-frozen granulocytes without technical challenges. Here, we propose to utilize buffy coat samples from 20 prostate, 20 colon, 20 breast, and 20 lung cancer patients together with 20 matched controls at three different time points in the PLCO cohort. We will generate single-cell NOMe-HiC data from these samples and identify the early-stage cancer-specific genomic regions, epigenetic marks, and relevant cell types.
Aim 2. Screen multiple early-stage cancers simultaneously by epigenetic marks inferred from cfDNA fragmentation.
The cfDNA fragmentation patterns are tightly correlated with the epigenetic marks in cells that contributed to them. In the preliminary studies, we developed computational tools to utilize cfDNA fragmentation patterns from low-coverage (~1X) WGS to predict different epigenetic marks, including open chromatin regions (Zhou et al 2022 Genome Med), DNA methylation (Liu et al. 2024 Nature Comm), and 3D genome (unpublished). We will generate low-coverage (~1X) cfDNA WGS from 200 prostate, 200 colon, 200 breast, and 200 lung cancer patients together with 200 matched controls at three different time points in the PLCO cohort. We will build a machine learning classifier by using inferred epigenetic marks from the genomic regions identified in Aim 1 to predict multiple early-stage cancers. We have already generated similar results from another relatively smaller cohort, which can be utilized to validate the model.

Collaborators

Yaping Liu (Northwestern University Feinberg School of Medicine)
Li Wang (Northwestern University Feinberg School of Medicine)