LONDON (March 22, 2022) InstaDeep today announced it will use the NVIDIA Cambridge-1 supercomputer to accelerate the next wave of biology innovation. The company plans to train AI language models using genomics data, and these models will then be shared with the bioinformatics community.
InstaDeep’s work on machine learning and bioinformatics has already led to the innovative DeepChainTM AI platform; and the new agreement for the $100 million Cambridge-1 will spur further research and development in the field. Breakthrough innovation as well as attention-based language transformer models make it possible to learn from raw biological sequences. This revolutionary approach to biology research complements and enhances current methods. An NVIDIA DGX SuperPOD, the powerful Cambridge-1 supercomputer will accelerate this research even further, enabling enterprise users to save time, lives and money in the drug discovery process.
Karim Beguir, co-founder and CEO of InstaDeep, said, “I am thrilled by this new collaboration with NVIDIA on Cambridge-1. It will enable us to significantly scale our platform to generate novel data-driven insights in genome sequencing, from over 12 billion nucleotide sequences, including non coding regions. AI models running on world-class infrastructure such as Cambridge-1 will allow us to process and interpret raw genetic information at an unprecedented scale, offering promising new avenues for biological research and development. We look forward to this new step in our collaboration with NVIDIA.”
Craig Rhodes, EMEA Industry Lead AI for Healthcare and Life Science at NVIDIA, added: “InstaDeep is a key software development partner deploying on NVIDIA DGX infrastructure in the world of digital biology. We are delighted that InstaDeep is leveraging Cambridge-1 to drive further breakthroughs in drug research and discovery.”
InstaDeep is one of a select few organisations granted access to Cambridge-1’s supercomputing capabilities, joining the five founding partners and start ups using it for projects in drug discovery, medical imaging and genomics – researching brain diseases like dementia, using AI to design new drugs, and more. AI-driven drug discovery workflows need the power of supercomputing for items such as simulation, virtual screening, protein structure prediction, medical imaging, NLP, and genomics.