Droplet Biosciences Cuts Cancer Genomic Analysis Time with NVIDIA GPUs

Droplet Biosciences has reduced its genomic analysis time for residual cancer detection by adopting NVIDIA's AI infrastructure, announced in a press release. The company's LymphDetect test, which identifies residual cancer in lymphatic fluid after tumor removal, now processes data significantly faster using NVIDIA Parabricks software.

By moving from CPU-based systems to GPU-accelerated workflows, Droplet shortened key bioinformatics steps substantially. Variant calling times dropped from up to 36 hours to under three hours, sequence alignment from around ten hours to under one, and overall analysis timelines from ten days to two days. These improvements are powered by NVIDIA's L4 Tensor Core and L40S GPUs.

Droplet noted that despite higher hourly GPU costs, the shorter runtime lowers the total cost per sample and supports scalability in clinical use. The company plans to further optimize its workflows with the goal of achieving one-day analysis for post-surgical cancer detection.