Arioc: High-concurrency short-read alignment on multiple GPUs

From GPUs to AI and quantum: three waves of acceleration in bioinformatics

The enormous growth in the amount of data generated by the life sciences is continuously shifting the field from model-driven science towards data-driven science. The need for efficient processing has led to the adoption of massively parallel accelerators such as graphics processing units (GPUs). Consequently, the development of bioinformatics methods nowadays often heavily depends on the effective use of these powerful technologies. Furthermore, progress in computational techniques and architectures continues to be highly dynamic, involving novel deep neural network models and artificial intelligence (AI) accelerators, and potentially quantum processing units in the future. These are expected to be disruptive for the life sciences as a whole and for drug discovery in particular. Here, we identify three waves of acceleration and their applications in a bioinformatics context: (i) GPU computing, (ii) AI and (iii) next-generation quantum computers.

BMX: Biological modelling and interface exchange

High performance computing has a great potential to provide a range of significant benefits for investigating biological systems. These systems often present large modelling problems with many coupled subsystems, such as when studying colonies of bacteria cells. The aim to understand cell colonies has generated substantial interest as they can have strong economic and societal impacts through their roles in in industrial bioreactors and complex community structures, called biofilms, found in clinical settings. Investigating these communities through realistic models can rapidly exceed the capabilities of current serial software. Here, we introduce BMX, a software system developed for the high performance modelling of large cell communities by utilising GPU acceleration. BMX builds upon the AMRex adaptive mesh refinement package to efficiently model cell colony formation under realistic laboratory conditions. Using simple test scenarios with varying nutrient availability, we show that BMX is capable of correctly reproducing observed behavior of bacterial colonies on realistic time scales demonstrating a potential application of high performance computing to colony modelling. The open source software is available from the zenodo repository https://doi.org/10.5281/zenodo.8084270 under the BSD-2-Clause licence.

Data-Rich Spatial Profiling of Cancer Tissue: Astronomy Informs Pathology

Astronomy was among the first disciplines to embrace Big Data and use it to characterize spatial relationships between stars and galaxies. Today, medicine, in particular pathology, has similar needs with regard to characterizing the spatial relationships between cells, with an emphasis on understanding the organization of the tumor microenvironment. In this article, we chronicle the emergence of data-intensive science through the development of the Sloan Digital Sky Survey and describe how analysis patterns and approaches similarly apply to multiplex immunofluorescence (mIF) pathology image exploration. The lessons learned from astronomy are detailed, and the new AstroPath platform that capitalizes on these learnings is described. AstroPath is being used to generate and display tumor-immune maps that can be used for mIF immuno-oncology biomarker development. The development of AstroPath as an open resource for visualizing and analyzing large-scale spatially resolved mIF datasets is underway, akin to how publicly available maps of the sky have been used by astronomers and citizen scientists alike. Associated technical, academic, and funding considerations, as well as extended future development for inclusion of spatial transcriptomics and application of artificial intelligence, are also addressed.

Performance optimization in DNA short-read alignment

SUMMARY

Over the past decade, short-read sequence alignment has become a mature technology. Optimized algorithms, careful software engineering and high-speed hardware have contributed to greatly increased throughput and accuracy. With these improvements, many opportunities for performance optimization have emerged. In this review, we examine three general-purpose short-read alignment tools-BWA-MEM, Bowtie 2 and Arioc-with a focus on performance optimization. We analyze the performance-related behavior of the algorithms and heuristics each tool implements, with the goal of arriving at practical methods of improving processing speed and accuracy. We indicate where an aligner's default behavior may result in suboptimal performance, explore the effects of computational constraints such as end-to-end mapping and alignment scoring threshold, and discuss sources of imprecision in the computation of alignment scores and mapping quality. With this perspective, we describe an approach to tuning short-read aligner performance to meet specific data-analysis and throughput requirements while avoiding potential inaccuracies in subsequent analysis of alignment results. Finally, we illustrate how this approach avoids easily overlooked pitfalls and leads to verifiable improvements in alignment speed and accuracy.

CONTACT

richard.wilton@jhu.edu.

SUPPLEMENTARY INFORMATION

Appendices referenced in this article are available at Bioinformatics online.

Accelerated genome sequencing with controlled costs for infants in intensive care units: a feasibility study in a French hospital network

Obtaining a rapid etiological diagnosis for infants with early-onset rare diseases remains a major challenge. These diseases often have a severe presentation and unknown prognosis, and the genetic causes are very heterogeneous. In a French hospital network, we assessed the feasibility of performing accelerated trio-genome sequencing (GS) with limited additional costs by integrating urgent requests into the routine workflow. In addition to evaluating our capacity for such an approach, this prospective multicentre pilot study was designed to identify pitfalls encountered during its implementation. Over 14 months, we included newborns and infants hospitalized in neonatal or paediatric intensive care units with probable genetic disease and in urgent need for etiological diagnosis to guide medical care. The duration of each step and the pitfalls were recorded. We analysed any deviation from the planned schedule and identified obstacles. Trio-GS was performed for 37 individuals, leading to a molecular diagnosis in 18/37 (49%), and 21/37 (57%) after reanalysis. Corrective measures and protocol adaptations resulted in a median duration of 42 days from blood sampling to report. Accelerated trio-GS is undeniably valuable for individuals in an urgent care context. Such a circuit should coexist with a rapid or ultra-rapid circuit, which, although more expensive, can be used in particularly urgent cases. The drop in GS costs should result in its generalized use for diagnostic purposes and lead to a reduction of the costs of rapid GS.

Serverless computing in omics data analysis and integration

A comprehensive analysis of omics data can require vast computational resources and access to varied data sources that must be integrated into complex, multi-step analysis pipelines. Execution of many such analyses can be accelerated by applying the cloud computing paradigm, which provides scalable resources for storing data of different types and parallelizing data analysis computations. Moreover, these resources can be reused for different multi-omics analysis scenarios. Traditionally, developers are required to manage a cloud platform’s underlying infrastructure, configuration, maintenance and capacity planning. The serverless computing paradigm simplifies these operations by automatically allocating and maintaining both servers and virtual machines, as required for analysis tasks. This paradigm offers highly parallel execution and high scalability without manual management of the underlying infrastructure, freeing developers to focus on operational logic. This paper reviews serverless solutions in bioinformatics and evaluates their usage in omics data analysis and integration. We start by reviewing the application of the cloud computing model to a multi-omics data analysis and exposing some shortcomings of the early approaches. We then introduce the serverless computing paradigm and show its applicability for performing an integrative analysis of multiple omics data sources in the context of the COVID-19 pandemic.

Hamming-shifting graph of genomic short reads: Efficient construction and its application for compression

Graphs such as de Bruijn graphs and OLC (overlap-layout-consensus) graphs have been widely adopted for the de novo assembly of genomic short reads. This work studies another important problem in the field: how graphs can be used for high-performance compression of the large-scale sequencing data. We present a novel graph definition named Hamming-Shifting graph to address this problem. The definition originates from the technological characteristics of next-generation sequencing machines, aiming to link all pairs of distinct reads that have a small Hamming distance or a small shifting offset or both. We compute multiple lexicographically minimal k-mers to index the reads for an efficient search of the weight-lightest edges, and we prove a very high probability of successfully detecting these edges. The resulted graph creates a full mutual reference of the reads to cascade a code-minimized transfer of every child-read for an optimal compression. We conducted compression experiments on the minimum spanning forest of this extremely sparse graph, and achieved a 10 − 30% more file size reduction compared to the best compression results using existing algorithms. As future work, the separation and connectivity degrees of these giant graphs can be used as economical measurements or protocols for quick quality assessment of wet-lab machines, for sufficiency control of genomic library preparation, and for accurate de novo genome assembly.

We present a novel graph-based algorithm to compress next-generation short sequencing reads. The novelty of the algorithm is attributed to a new definition of genomic sequence graph named Hamming-Shifting graph. It consists of edges between distinct reads that have a small Hamming distance or a small shifting offset or both. Efficient construction of Hamming-Shifting graphs is challenging. We introduce a heuristic technique to detect the weight-lightest edges through multiple minimizers from each read, then search the minimum spanning trees and forests of the Hamming-Shifting graph for a high-performance compression of the reads. Our method achieves an additional 10 − 30% file size reduction compared to contemporary compression techniques.