Due to significant improvements in genomic sequencing technology, it is now possible to sequence a full human genome for less than $1000 and receive genetic data on several molecular levels of the cell (genomic, transcriptome, protein expression, copy-number variations, and so on). However, in order to turn this genomic Big Data into useful insights for precision medicine, a systematic integrated study of a range of biological datasets is required. ACCELOM develops bespoke, reproducible multi-omics software pipelines that combine numerous molecular datasets from a host to investigate association patterns between molecular levels using unique, peer-reviewed statistical machine learning algorithms.

ACCELOM pipelines rely on the well-established GATK pipeline for performing secondary analysis, which transforms raw sequenced reads into analysis-ready variants. It is well known that the GATK pipeline is extremely computationally intensive[1] especially for large datasets (ex. Whole Genome Sequencing). To address this bottleneck, several hardware acceleration solutions have emerged to scale secondary analysis[2] [3]. While these solutions often improve time-to-insight by several orders of magnitude, they are closed and proprietary on both the software front (as accelerated pipelines not freely available) and hardware front (as they are tied to a specific vendor hardware).

Through SYCLOPS, ACCELOM intends to collaborate with various partners in the development and validation of a SYCL-based library (SYCL-GAL) that will accelerate key stages of the GATK secondary analysis pipeline. Unlike contemporary solutions, SYCL-GAL library will be open source in nature to foster the adoption of open standards in genomics. Further, by relying on SYCL for cross-architecture parallelism, SYCL-GAL will also enable cross-platform acceleration of genomic data analysis, making it possible to deploy the same pipeline on CPUs, GPUs, and accelerators from several hardware vendors.