A cloud-based system for scalable, privacy-preserving, and interactive immune repertoire analysis in vaccination and infection

项目摘要 我们建议构建一个基于云的集成解决方案，以实现可扩展、可定制、隐私保护和 交互式抗体库分析。免疫库测序（IR-seq）已成为一个有用的工具， 基础研究和临床环境。作为感染和许多疫苗的适应性免疫的心脏， B细胞受体（BCR）的丰度和多样性组成及其在健康和 疾病承载着如何评估免疫健康、进行疾病诊断和预后的信息， 衡量疫苗接种效果。然而，大规模抗体谱系的计算存在一个计算瓶颈 建设，缺乏可分解的管道模块，保护隐私和所有权，以及缺失的差距 用于交互式图像分析和可视化。在第一期资助中，我们将（1）打破管道瓶颈 处理和扩展核心算法以处理大型序列数据集;（2）保护私有数据， 专有处理算法，具有模块化管道、集成云本地处理和数据 扰动方法; 3）开发用于管道组合和交互分析的端到端Web服务 在基于云的部署解决方案中实现可视化。现有的商业努力主要集中在癌症上 相关的IR-seq分析，旨在追踪治疗后癌细胞的消失，其仅关注 分类序列物种和丰度。这样的分析要简单容易得多， 分析感染和疫苗接种中的IR-seq数据。提供宿主免疫反应的见解是一个更重要的 具有挑战性，但非常需要的任务。一旦管道建成，它可以很容易地适用于分析癌症IR-seq 数据此外，已经为其他大数据分析开发的现有算法和优化可以被 进一步开发并应用于IR-seq数据分析。我们将使用公开的BCR库数据 关于流感疫苗接种队列和关于老龄化队列的TCR库数据，以测试 项目该提案的长期目标是构建基于云的可访问和可定制服务， 专家和非专家。我们的目标是提供一个综合解决方案的摄入，处理， 分析，探索和可视化，解释和共享由深度测序产生的数据， 长度抗体和TCR库。第一阶段SBIR的成功将为 在第二阶段推出基于云的商业解决方案，在此期间，我们将继续调查 大数据安全和隐私，以促进遵守机构政策，并设计和开发 可编程API和工具，以促进与更多第三方模块和服务的集成。

PROJECT SUMMARY We propose to build a cloud-based integrated solution for scalable, customizable, privacy-preserving, and interactive antibody repertoire analysis. Immune repertoire sequencing (IR-seq) has become a useful tool in both basic research and clinical settings. As the heart of the adaptive immunity to infection and many vaccines, the abundance and diversity composition of the B cell receptor (BCR) and its dynamic changes in health and diseases bear information of how to evaluate immune health, perform disease diagnosis and prognosis, and measure vaccination effect. However, there is a computational bottleneck for large scale antibody lineage construction, a lack of decomposable pipeline modules that preserve privacy and ownership, and a missing gap for interactive linage analysis and visualization. In this Phase I grant, we will (1) break the bottlenecks of pipeline processing and scale up the core algorithms to handle large sequence data sets; (2) protect private data and proprietary processing algorithms with modularized pipeline, integrated cloud-local processing, and data perturbation methods; 3) develop end-to-end web services for pipeline composition and interactive analysis visualization in a cloud-based deployment solution. Existing commercial efforts are mostly focusing on cancer related IR-seq analysis aiming to trace the disappearing of cancer cells after therapy, which solely focus on cataloging sequence species and abundance. This kind of analysis is much simpler and easier, compared to analyzing IR-seq data in infection and vaccination. Providing insights on host immune responses is a much more challenging but much needed task. Once the pipeline is built, it can be readily adapted to analyze cancer IR-seq data. Also, existing algorithms and optimizations that have been developed for other big data analysis can be further developed and applied to the IR-seq data analysis. We will use a publically available BCR repertoire data on an influenza vaccination cohort and a TCR repertoire data on an aging cohort to test the feasibility of the project. The long term goal of this proposal is to build cloud based accessible and customizable services for experts as well as non-specialists. We aim to provide an integrated solution for the ingestion, processing, analysis, exploration and visualization, interpretation and sharing of data generated by deep sequencing of full length antibody and TCR repertoire. The success of this Phase I SBIR will provide a solid foundation for the product launch of a commercial cloud based solution in Phase II, during which we will continue our investigation on big data security and privacy to facilitate compliance with institutional policies and design and develop programmable APIs and tools to facilitate integrations with more third party modules and services.