An external research seminar was virtually held at the MRC Institute of Metabolic Research, fostering discussion on integrating AI technology, HPC infrastructures, and Scientific Pipelines.
On February 12th, Prof. Marco Aldinucci of the University of Torino presented the seminar “Pipeline-as-a-Service with Streamflow” hosted by the IMS-MRL Bioinformatics and Biostatistics Core (BIO2) directed by Dr. Davide Chiarugi.
The HPC (High Performance Computing) systems evolution and their ability to calculate many mathematical operations per second (FLOPS) has been the key to solving increasingly larger problems and finding ever more precise solutions. The explosion of data resulting from digital transformation has shifted the demand for HPC from traditional applications to methods for analyzing large amounts of data, from toggling absolute binary in the front panel of the machine to an ecosystem of high-level libraries and services on the cloud.
In the seminar, Prof. Aldinucci depicted the current technological landscape and presented Streamflow, a novel Workflow Management System developed in the DeepHealth project. StreamFlow is based on the CWL open standard to support the transition from application pipelines to modern, dockerized, portable, R/Jupyter-enabled Pipeline-as-Service for hybrid HPC-cloud systems (e.g., SLURM, K8S).
He also presented OpenDeepHealth, a secure HPC multitenancy infrastructure developed in DeepHealth, and further applications examples like single-cell pipelines, QuantumExpresso-as-a-service, the digital twins for seismic (SPECFEM3D, SPEED, OPENSEES), the universal pipelined for AI-enabled diagnosis of COVID-19 interstitia.
The interest raised by the presentation and the following discussion on the possibility of applying DeepHealth technology to bioinformatics pipelines were the first steps to start a new collaboration with the hosting laboratory.