Software is a key component of deep-sequencing and point-of-care tools. Smart software solutions allow to speed up the sample-to-answer time, and to save on the system’s power. The importance of software – in the form of machine learning algorithms and artificial intelligence – will only become more prominent in the future when the amounts of data will explode due to the emergence of proteomics, metabolomics, single-cell sequencing, etc. and when millions of tests will be performed in the daily practice of hospitals worldwide as the cornerstone of their personalized medicine practice.
Imec has set up a lab, the ExaScience Life Lab, that focuses on software solutions for data-intensive and high-performance computing problems in life sciences. As part of that endeavour the lab developed open-source tools for speeding up the DNA reconstruction and variant calling process as well as very fast single-cell RNAseq software.
Another challenge that the lab is tackling is privacy- and ownership-preserving solutions for machine learning and genomic analysis. In this webinar, we will explain how such amalgamated machine learning could enable a form of federated analytics across multiple parties such that each party keeps its own data and analytics private yet benefits from others by only sharing non-sensitive derived model information.
Roel Wuyts leads the ExaScience Life Lab, a lab focused on providing software solutions for data-intensive high-performance computing problems, primarily in (but not limited to) the life sciences domain.
The lab has extensive experience with high performance computing technologies (distributed computing, parallel computing, concurrent computing, vectorization, NUMA optimizations), programming languages (Go, C++, Python, Lua, Rust, and many more), and usage of hardware accelerators (GPU, Xeon Phi, FPGA). It helps companies develop prototype solutions for complex problems involving multiple disciplines, thereby applying the experience to remove computational bottlenecks. The lab scientists have successfully done this in the past for large-scale machine learning for pharmaceutical companies and imec projects, DNA sequencing software for hospitals and pharmaceutical companies, assay image feature extraction, advanced biostatistics and data analytics, and even multi-physics space weather simulations.
Roel Wuyts is also part-time professor in the Distrinet group at the KU Leuven. He published papers in IEEE Software or the Journal of Systems and Software, TOPLAS, ECOOP, OOPSLA or AOSD.