BC Platforms enhances data analytics capabilities with NVIDIA accelerated computing 

NVIDIA technology now natively available in BC Mosaic, our trusted research environment

We recently announced a series of enhancements to BC Mosaic, our trusted research environment (TRE), including the addition of high-performance computing architecture using NVIDIA software and computing technologies. BC Mosaic provides healthcare and life sciences researchers with direct and out-of-the-box access to NVIDIA GPUs, NVIDIA Rapids, and other optimized NVIDIA data science frameworks.  

This integration will enable faster and more scalable GPU-powered analytics, as well as development of new AI-led analytics across clinical, multi-omics and real-world data, to help life sciences and healthcare organizations advance precision medicine and increase the speed – and success – of drug innovation. 

Enabling 80x faster AI-driven analytics across multi-modal data

BC Mosaic will be able to process complex multi-modal datasets as much as 80 times faster than with traditional computing architectures while also supporting new analysis of old data. Researchers working in BC Mosaic can apply powerful AI and machine learning to federated datasets without moving or exposing sensitive information, thereby protecting data privacy and complying with data governance standards. 

“To develop novel therapies and advance precision medicine, the industry needs greater speed, broader scale, and advanced analysis of complex health data,” says Keith Collier, Chief Product Officer at BC Platforms. “By integrating NVIDIA technologies into our trusted research environments, BC Platforms can support sophisticated analyses at a pace and magnitude that were not previously possible. Our workflow simulations show performance improvements that are 80 times faster compared to traditional processing times – truly significant gains. This helps life sciences companies, biobanks, and healthcare organizations more efficiently unlock novel insights from real-world and multi-omics data to better inform discovery, development, and commercialization decisions.”

Validating performance across real-world research workflows 

Simulations were conducted across large scale, quantifiable trait datasets using NVIDIA T4 GPUs on AWS EC2 instances (G4). The simulations were focused on population stratification workflows using cuGRAPH and NetworkX for Eigenvector analysis.

Additionally, cuDF and cuGraph library methods were used to simulate community clustering analysis. An 80 times faster performance improvement in cuDF-based data processing and algorithms was observed in comparison to traditional pandas-based CPU-only analysis. These use cases are typically conducted by healthcare institutions and biopharma research labs as part of their early-stage R&D and clinical trial analyses.  

Future collaboration efforts will see these NVIDIA tools and frameworks utilized in BC Mosaic for further testing and use by leading national healthcare institutes. 

BC Image: High-volume medical imaging also powered by NVIDIA

NVIDIA technology already powers BC Image, which automates high-volume image extraction, de-identification, and processing. 

BC Image is widely used in hospitals and treatment centers across Europe, where it has the capacity to de-identify and harmonize millions of images a week and make them available to support advanced research and analyses. 

NVIDIA CUDA and other NVIDIA tools enable BC Image to augment medical images with valuable contextual data, granular information, and filter tags that increase the speed and efficiency of subsequent analysis.