Simple and Accessible Microfluidic Platform for Single Molecule Sequence Profiling of Tumor-derived DNA within Liquid Biopsies

项目摘要/摘要 今年，美国将有超过60万人死于癌症。据估计，这些死亡中有25%可能 已经通过在早期阶段进行检测来预防。实施微创常规筛查， 例如子宫颈癌的宫颈涂片，已被证明是降低癌症死亡率的有效方法。 然而，有几个挑战阻碍了在大多数癌症，特别是癌症中成功实施筛查 它们不容易被用于成像或组织活检。癌症诊断的一个有希望的途径是 通过使用来自所谓的“液体活组织检查”或其他可获得的循环样本介质的循环DNA 从包括肿瘤在内的组织中收集遗传物质。肿瘤特异性分子 生物标记物，如DNA突变和甲基化，可以在微创样本培养物中找到，如 作为血液、粪便和尿液，但通常只存在于非常低的拷贝数(&lt；10拷贝/毫升)和低拷贝数 在健康DNA的高背景中的部分(&lt；0.1%)。技术限制和实用 目前可用的工具无法获得，这阻碍了发现早期癌症的研究努力-- 特定的DNA生物标记物面板和随后的临床实施可以改善患者的预后。 为了解决这个问题，我们之前开发了一个原型数字微流控平台，以促进高度敏感的， 用高度平行化的单分子低成本检测癌症特异性DNA甲基化模式 热力学测序。在此阶段1 SBIR提案中，我们将极大地扩展 增加可访问性并改进早期疾病检测的分析性能的平台 通过大幅提高其数字化能力。我们将开发一个高度多路复用的范例， 使用多色条形码技术对生物标志物面板进行检测和甲基化分析。到时候我们会的 将这种分析整合到微流控平台中，该平台可以询问数百到数千个单一DNA 通过将模板分子数字化成液滴进行高通量单分子分析。这个 平台将通过将成本降低到每英镑25美元，增加液体活组织检查的生物标记物研究的可及性 样品和周转时间为4小时。该平台将支持单拷贝检测，即使在高 背景人群(0.001%的敏感性)，这可能是早期疾病所必需的。建议数 这一阶段项目的工作将为多维、多维的分析奠定基础 临床相关生物标志物小组(AIM 1)，将此检测纳入高度并行的液滴微流控技术 平台(Aim 2)，并用一组肺癌患者的血浆样本评估其临床可行性 和对照(目标3)。这将为随后的第二阶段项目设计弹药筒奠定基础 和仪器，以实现可扩展的制造和用户友好性，并实施自动化、机器- 学习图像和数据分析管道。

Project Summary/ Abstract Over 600,000 people in the US will die from cancer this year. It is estimated that 25% of these deaths could have been prevented by detection in earlier stages. Implementation of minimally-invasive routine screening, such as pap smears for cervical cancer, has proven to be an effective approach for reducing cancer mortality. However, several challenges prevent successful implementation of screening in most cancers, especially ones which are not readily accessible for imaging or tissue biopsy. One promising avenue for cancer diagnostics is through use of circulating DNA from so-called “liquid biopsies” or other accessible sample media circulating throughout the body collecting genetic material from tissues, including tumors. Tumor-specific molecular biomarkers, such as DNA mutations and methylation, can be found in minimally-invasive sample media, such as blood, stool, and urine, but are typically only present in very low copy numbers (<10 copies/mL) and low fractions (<0.1%) among a high background of healthy DNA. Technical limitations as well as practical inaccessibility of currently available tools have precluded research efforts to discover early-stage cancer- specific DNA biomarker panels and subsequent clinical implementation that could improve patient outcomes. To address this, we previously developed a prototype digital microfluidic platform to facilitate highly sensitive, low-cost detection of cancer-specific DNA methylation patterns by highly parallelized single-molecule thermodynamic sequencing. In this Phase 1 SBIR proposal, we will greatly expand upon the capabilities of this platform to increase accessibility and improve analytical performance towards detection of early-stage disease by significantly increasing its digitization power. We will develop a high-degree multiplexing paradigm for detection and methylation profiling of biomarker panels using a multicolor barcoding technique. We will then incorporate this assay into a microfluidic platform that can interrogate hundreds to thousands of single DNA copies by digitizing template molecules into droplets for high-throughput single molecules analysis. The platform will increase accessibility for biomarker research from liquid biopsies by reducing costs to <$25 per sample and turnaround time to 4 hours. The platform will enable single-copy detection even among high background populations (<0.001% sensitivity), which may be necessary for early-stage disease. The proposed work in this Phase 1 project will develop the assay fundamentals for a multiplex, multidimensional analysis of a clinically relevant biomarker panel (Aim 1), incorporate this assay into a highly-parallelized droplet microfluidic platform (Aim 2), and assess its clinical feasibility with plasma samples from a cohort of lung cancer patients and controls (Aim 3). This will lay the groundwork for a subsequent Phase 2 project to design the cartridge and instrumentation for scalable manufacturing and user-friendliness and implement automated, machine- learning image and data analysis pipelines.