Biomarker discovery, high performance and cloud computing: A comprehensive review

Abstract

Abstract-The analysis of biological markers (biomarkers) have the ability to improve clinical outcomes of a disease through prediction and early detection. The constant improvement of next generation sequencing (NGS) technologies coupled with falling equipment price are driving research and application especially in the medical domain. NGS technology is being applied to cancer research promising greater understanding of carcinogenesis. However, as sequence capacity grows, algorithmic speed is becoming an important bottleneck. To understand these challenges we present a review on difficulties currently faced in discovering biomarkers. The review places the spotlight on sequencing technologies and the bottlenecks encountered in these pipelines. Cloud computing and high performance computing technologies such as Grid are summarized in tackling computational challenges presented by technologies and algorithms used in biomarker research.

Review of applications of high-throughput sequencing in personalized medicine: barriers and facilitators of future progress in research and clinical application

“…Commodity clusters (Figure 3, Supplementary Table S1) have attained popularity within bioinformatics, because of their relatively low cost and scalability [152, 153]. They consist of regular desktops, with central processing units (CPUs) (for handling computations) or networked with servers (larger versions of desktops), [154] linked together to form a distributed computer system. This type of infrastructure enables parallel computing to be undertaken in (small) laboratories using low-cost hardware and standard software.…”

“…GPUs (Figure 3, Supplementary Table S2) are card-based devices, which can be slotted into the graphics port of a laptop or desktop. One GPU card can comprise hundreds of computational units, in comparison with a CPU, offering increased scalability and processing performance [154, 181, 182]. Considering the price to performance ratio, parallel GPUs are potentially a more affordable and efficient option when compared with multiple, sequential CPUs [181–183].…”

Custom Hardware Versus Cloud Computing in Big Data

“…They enable highly mathematical and computationally intense functions to be performed at an accelerated rate due to the parallel computational units at the heart of their structure. The ability to offload computation most suited to parallel operations, while maintaining a great level of flexibility and scalability is a leading benefit of GPU-based computing over sequential operation CPU-based computing [Blayney et al, 2015, Melanakos, 2008, Fan et al, 2004. However, the scale of the benefits depends strongly on the nature of the computations.…”

The role of high performance, grid and cloud computing in high-throughput sequencing

“…Commodity clusters consist of regular computers (servers) which are connected through the network as compared to a supercomputer with many processors [29]. Here, a regular computer may also have multiple processing cores.…”

“…Many researchers [29], [40], [41], compare Graphics Processor Unit (GPU)-based computing with a traditional CPU-based parallel computing. According to the price to performance ratio, parallel in nature GPUs are potentially more affordable and efficient as compared to sequential in nature CPUs [41], [40].…”