CRII: SHF: HPC Solutions to Big NGS Data Compression

由于下一代高通量基因组测序(NGS)技术，包括人类在内的众多物种的基因组测序已经变得越来越负担得起。这为遗传病的诊断和治疗开辟了前景，并在进行系统生物学研究方面日益有效。然而，在这些技术进入日常健康和人类护理之前，仍然有许多计算挑战需要解决。其中一个令人望而生畏的挑战是测序数据的数量，这些数据可以达到Peta字节级别，用于全面的系统生物学研究。需要对基因组数据进行压缩，以减小存储大小、提高传输此类数据所需的I/O带宽的速度和降低成本。然而，现有的基因组压缩解决方案对大基因组数据的性能很差。此外，现有技术水平的工具要求用户在将数据用于进一步分析之前对其进行解压缩。该项目的重点是压缩基因组信息，并开发一个框架，以便对数据的压缩形式进行分析。该项目使用无处不在的架构，如GPU和多核处理器，开发用于快速压缩大NGS数据集的高性能计算解决方案。利用HPC技术来计算基本函数，例如使用NGS数据的压缩形式的对准和映射。为了更好地利用网络，还在研究对NGS数据进行更有效的编码。

Sequencing of genomes for numerous species including humans has become increasingly affordable due to next generation high-throughput genome sequencing (NGS) technologies. This opens up perspectives for diagnosis and treatment of genetic diseases and is increasingly effective in conducting system biology studies. However, there remain many computational challenges that need to be addressed before these technologies find their way into every day health and human care. One such daunting challenge is the volume of sequencing data which can reach peta-byte level for comprehensive system-biology studies. Genomic data compression is needed to reduce the storage size, to increase the speed and reduce the cost of I/O bandwidth required for transmission of such data. However, existing genomic compression solutions yield poor performance for Big Genomic Data. Further, the existing state of the art tools require the user to decompress the data before it can be used for further analysis. This project is focused on compression of genomic information and developing a framework which will allow analysis of compressed form of the data. The project develops HPC solutions for fast compression of Big NGS Data sets using ubiquitous architectures such as GPUs and multicore processors. HPC techniques are utilized to compute essential functions such as alignment and mapping using the compressed form of the NGS data. More efficient encoding of the NGS data for better network utilization is also being investigated.