Fragmentomic and Epigenetic Analyses for Cell-free DNA Molecules

Circulating cell-free DNA (cfDNA) molecules released from various tissues within the body are nonrandomly fragmented. There is increasing research interest in studying the nascent field of fragmentomics of cfDNA in plasma. Fragmentomics of cfDNA includes information regarding fragment sizes, nucleosome footprints, end-point positionings, and end motifs. In this thesis, I developed approaches to uncover new properties of cfDNA and elucidate their potential clinical utilities on the basis of fragmentomic and epigenetic analyses.In the 1st chapter, I developed genetic-epigenetic tissue mapping (GETMap) to determine the tissues of origin for plasma DNA carrying variants not present in the background DNA to be mainly of hematopoietic origin. GETMap was implemented by comparing the methylation patterns of plasma DNA with variant alleles with the counterpart of relevant tissues. This approach had been validated by the fact that circulating cfDNA carrying fetal-specific variants was determined to be entire of placental origin. I subsequently applied this approach to analyze patients with lung transplants. It was found that the lungs contributed only a median of 17% to the plasma DNA carrying donor-specific alleles, while the hematopoietic cells contributed up to a median of 78% at 72 hours after transplantation. This suggested that a substantial contribution of DNA from the donor genome in the plasma DNA pool was derived from the hematopoietic cells after the operation, instead of having a graft origin. For patients with hepatocellular carcinoma, the plasma DNA carrying variants absent in the patient's constitutional genomes was found to originate mainly in the liver. This new technology for tissues of origin analysis for plasma DNA carrying genetic variants is potentially useful for noninvasive prenatal testing (NIPT), transplantation monitoring, and cancer detection.In the 2nd chapter, I developed a new approach to perform the deconvolution of tissues of origin of plasma DNA on the basis of plasma DNA sizes. Our group previously demonstrated that fetal DNA molecules were shorter than their maternal counterparts, whose size profile showed 143- and 166-bp major peaks with a series of smaller peaks with 10-bp intervals. However, such a 143-bp peak appeared to be absent in the maternal DNA molecules. Thus, I hypothesized that cfDNA released from different tissues would have different size characteristics. The mixture of cfDNA would be deconvoluted on the basis of the characteristic size profiles, allowing for the tracing of the tissues of origin of plasma DNA. In this chapter, I analyzed the plasma DNA of pregnant samples and cancer subjects to demonstrate the feasibility of size-based plasma DNA tissue mapping.In the 3rd chapter, I analyzed the end motifs of circulating mitochondrial DNA (mtDNA) between colorectal cancer (CRC) patients with and without liver metastases. I found that the patterns of end motifs of circulating mtDNA could provide a tool to determine whether a CRC patient would be subjected to liver metastases.In the 4th chapter, I developed methylation-based approaches to reveal the existence of jagged ends of urinary DNA. Jagged ends were herein referred to as the single-stranded protruding ends carried by double-stranded plasma DNA molecules. These approaches relied on the DNA end- repair step to introduce differential methylation signatures between the original double-stranded DNA and the jagged ends, named Jag-seq. For example, the incorporated unmethylated cytosines due to the end repair of jagged ends would decrease the methylation levels in positions close to 3' ends. The degree of methylation decrease between read 1 and read 2 allowed for the quantification of the jaggedness of plasma DNA. In this study, it was revealed that the jaggedness varied according to plasma DNA fragment sizes, displaying multiple nucleosomal patterns. The urinary DNA molecules were found to have higher jaggedness than the plasma DNA in paired healthy subjects. Interestingly, a lower jaggedness level was present in bladder tumor-derived DNA molecules compared with background urinary nontumoral DNA. I further demonstrated that nuclease activities were responsible for the generation of plasma DNA jagged ends based on gene expression-level changes in higher-grade bladder cancer. This work illustrated an intrinsic property of urinary DNA and provided a biological link between nucleases and the fragmentation of urinary DNA, which could open up new possibilities for liquid biopsy.In summary, I developed various bioinformatics approaches for revealing the biological properties of cfDNA molecules, thus opening up new opportunities for liquid biopsy for NIPT, organ transplantation, and cancer detection.人體中的循環游離 DNA 是由多種組織釋放而出進入血液的非隨機斷裂所形成的核酸片段。 如今，游離 DNA 片段組學作為一個新興的研究領域而備受矚目。游離 DNA 片段組學包含 豐富的訊息，諸如片段長度、核小體印記、末端位置和末端序列模序等。在本篇論文之 中，我開發出多種創新性研究方法用以揭示游離 DNA 新的生物學特性，並且闡明如何基於 片段組學和表觀遺傳學建立起潛在的臨床應用。在本文第一章节中，我建立了一個用於游離 DNA 的遺傳學-表觀遺傳學的組織溯源方法 (GETMap)，並且依此確定血漿游離 DNA 中帶有變異的分子主要來自於造血細胞。這個 方法將攜帶突變的游離 DNA 和相應器官組織的甲基化特徵進行比較來實現組織溯源。利用 這一方法，我驗證了胎兒游離 DNA 完全來自於胎盤組織這一事實。隨後我將這一分析方法 用於接受肺移植的病人當中。肺移植手術 72 小時內，在移植器官所釋放的游離 DNA 中， 肺組織只占約 17%;然而造血細胞卻貢獻 DNA 總數的約 78%。這一結果表明，在器官移植 的初期，移植器官所釋放的游離 DNA 中有相当可觀的比例來自於造血細胞而非被移植器官 本身。對於肝癌患者而言，血漿游離 DNA 常帶有變異，此組織溯源方法將這些與患者本身 基因組相左的變異定位到肝臟組織。這個新方法可以對游離 DNA 中帶有變異的 DNA 分子 進行組織溯源分析，可能可以用於無創傷產前檢測、移植監測和癌症診斷。在本文第二章节中，我建立了新的血漿游離 DNA 器官組織溯源方法，即基於 DNA 片段長 度進行反卷積分析。在此之前，我們團隊已經表明相較於母體 DNA，胎兒游離 DNA 長度 分佈更短，同時胎兒游離 DNA 呈現出 143-bp 和 166-bp 的主峰和多個擁有 10-bp 間隔的次 峰。並且位於 143-bp 的峰在母體游離 DNA 長度分佈中並不明顯。於是我假設不同組織釋放的游離 DNA 片段會有不同的長度分佈。如果基於游離 DNA 長度分佈特徵進行反卷積分 析，將會達到追溯游離 DNA 組織來源的目的。在本文這一章节中，我通過對於懷孕個體和 癌症病患的游離 DNA 進行分析研究，證實基於片段長度的血漿游離 DNA 組織溯源分析的 可行性。在本文第三章节中，我分析了線粒體游離 DNA 末端模序序列在結直腸癌患者的肝臟轉移檢 測中的應用。我發現線粒體游離 DNA 末端序列特徵模序可作為一個臨床檢測方法，用以鑑 定結直腸癌患者是否存在癌細胞肝臟轉移。本文第四章节中，我基於 DNA 甲基化建立起新的研究方法，並確定了尿液游離 DNA 存在 突出末端。突出末端這裡具體指:擁有雙鍊形态的血漿游離 DNA 分子以單鏈突出的型態作 為末端。用於研究這一現象的方法稱為 Jag-seq。具體而言，DNA 末端修復步驟會在原有序 列和突出末端引入不同的甲基化信號。譬如在修復位於末尾的單鏈突出末端時引入未甲基 化的胞嘧啶，藉此降低 3'末端序列的甲基化水平。進而導致測序產生的 read 1 與 read 2 其 甲基化水平將出現差異，而這一差異將用於量化分析突出末端的突出程度。這一研究揭示 了尿液游離 DNA 包含突出末端。並且突出末端的突出程度會隨著游離 DNA 長度變化而展 示出與核小體同步的週期性變化。尿液中的游離 DNA 相比血浆游離 DNA 包含更多的突出 末端。在膀胱腫瘤患者中，相比於其尿液的非腫瘤游離 DNA，腫瘤組織所釋放的游離 DNA 分子包含更少的突出末端。借助於基因敲除的小鼠，我進一步闡明了核酸酶活性會影 響尿液游離 DNA 中突出末端的產生。這一研究展示出了游離 DNA 的一種內在特性，並且 將核酸酶與游離 DNA 片段化過程從生物學角度聯繫起來，為新的液體活檢方法提供依據。綜上所述，我開發了多種生物訊息學研究方法，揭示了游離 DNA 分子的部分生物學性質，並為無創傷產前檢測、器官移植監測和癌症發現等液體活檢應用開拓了諸多可能。ZHOU, Ze.Ph.D. Chinese University of Hong Kong 2021.Includes bibliographical references (leaves )Abstracts in English and Chinese.Title from PDF title page (viewed on ...