EAGER: High-throughput early detection and analysis of COVID-19 plaque formation using time-lapse coherent imaging and deep learning

EAGER：使用延时相干成像和深度学习对 COVID-19 斑块形成​​进行高通量早期检测和分析

• 批准号: 2034234
• 负责人: Aydogan Ozcan
• 金额: $ 29.99万
• 依托单位: University of California-Los Angeles
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-15 至 2023-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2034234&HistoricalAwards=false
• 关键词: EAGER; throughput; early; detection; analysis

Plaque assays are widely used for measuring the infectious concentration of viral samples and form a very important tool for vaccine development, especially for the evaluation of the performance of new vaccines at the exploratory and preclinical stages. This standard method is laborious and takes days to get the results, and is subject to human errors since it depends on manual plaque counting. Molecular techniques such as polymerase chain reaction (PCR or reverse transcription PCR) and western blots can be used to quantify the viral genome. However, none of these methods provide information about the infectivity of the virus and cannot measure plaque forming units. This proposal aims to create a computational sensor platform for accelerated testing of SARSCoV-2 viability and infectivity using deep learning-based plaque assays and achieve accurate and automated plaque forming unit (PFU) measurements within hours as opposed to days with standard plaque assays. The proposed computational imaging system will periodically capture coherent microscopic images of the cytopathogenic effects of viruses on cell cultures and analyze these time lapsed holographic images using deep neural networks (DNNs) for rapid detection of viral destruction of the cell monolayer. In addition to early and automated detection of plaque forming units, this unique platform will further make use of deep learning for high-throughput holographic image reconstruction of the assay volume to perform tile-scan imaging of the entire well plate within 5 min, corresponding to an imaging throughput of ~50 cm2/min. Powered by deep learning, this automated and cost-effective viral plaque monitoring platform can be transformative for a wide range of applications in microbiology and virology by significantly reducing the detection time without labeling or the need for an expert, or manual inspection. The project will also establish a complementary educational outreach program that will involve (1) public interviews and popular science articles in news media and internet; (2) undergraduate research opportunities in the PI’s laboratory involving minority students; and (3) graduate student training through organization of workshops, seminars and conferences. Furthermore, research projects, seminars and open house visits will serve undergrads and high school students (especially from minority groups) to interact with a cutting edge research environment, helping to increase their scientific curiosity and shaping their career goals in science and engineering.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

空斑试验广泛用于测量病毒样品的感染浓度，并形成疫苗开发的非常重要的工具，特别是用于评价新疫苗在探索和临床前阶段的性能。这种标准方法是费力的，需要几天时间才能得到结果，并且由于它依赖于人工噬菌斑计数，因此容易出现人为错误。分子技术如聚合酶链反应（PCR或逆转录PCR）和蛋白质印迹可用于定量病毒基因组。然而，这些方法都不能提供有关病毒感染性的信息，也不能测量空斑形成单位。该提案旨在创建一个计算传感器平台，用于使用基于深度学习的空斑试验加速SARSCoV-2活力和感染性的测试，并在数小时内实现准确和自动化的空斑形成单位（PFU）测量，而不是使用标准空斑试验的数天。所提出的计算成像系统将定期捕获病毒对细胞培养物的致细胞病变作用的相干显微图像，并使用深度神经网络（DNN）分析这些时间流逝的全息图像，以快速检测细胞单层的病毒破坏。除了早期和自动化检测噬菌斑形成单位外，这一独特的平台还将进一步利用深度学习进行分析体积的高通量全息图像重建，在5分钟内对整个孔板进行平铺扫描成像，对应的成像通量约为50 cm 2/min。这种自动化和成本有效的病毒空斑监测平台可以通过显著减少检测时间而无需标记或需要专家或人工检查来改变微生物学和病毒学中的广泛应用。 该项目还将建立一个补充的教育外展计划，其中包括：（1）新闻媒体和互联网上的公开采访和科普文章;（2）在PI实验室中涉及少数族裔学生的本科生研究机会;（3）通过组织研讨会、研讨会和会议对研究生进行培训。此外，研究项目，研讨会和开放参观将服务于本科生和高中生（特别是少数群体）与前沿研究环境互动，帮助提高他们的科学好奇心并塑造他们在科学和工程领域的职业目标。该奖项反映了NSF的法定使命，并通过利用基金会的知识价值和更广泛的影响进行评估，被认为值得支持审查标准。

项目成果

Stain-free, rapid, and automated viral plaque assay using time-lapse holographic imaging and deep learning

使用延时全息成像和深度学习进行无染色、快速、自动化的病毒斑块测定

• DOI: 10.1364/fio.2023.fm6e.2
• 发表时间: 2023
• 作者: Li, Yuzhu;Liu, Tairan;Koydemir, Hatice Ceylan;Zhang, Yijie;Yang, Ethan;Eryilmaz, Merve;Wang, Hongda;Li, Jingxi;Bai, Bijie;Ma, Guangdong
• 通讯作者: Ma, Guangdong