How Can AI Accelerate Drug Discovery and Development?

TL;DR

AI can significantly accelerate drug discovery by utilizing breakthroughs in computer science, such as in-silico methods and simulation, to streamline the development process. NVIDIA is introducing tools like Clara Discovery and building the Cambridge-1 computing infrastructure to bolster AI and healthcare collaborations, enabling faster identification of drug candidates and efficient research partnerships across the pharmaceutical industry.

Transcript

Understanding disease is one of the great human endeavors and remains an incredible challenge. Drug discovery is hard, on average taking over a decade to develop, $2.5 billion dollars in R&D. And yet 90% of the efforts fail, the cost has doubled every 9 years. This is called Eroom's law. For the $1.5 trillion dollar pharma industry there is a gre... Read More

Key Insights

• 🛩️ Drug discovery is a complex and challenging process, requiring a deep understanding of proteins, small molecules, cell penetration, and pharmaco-kinetics.
• 💻 Breakthroughs in computer science, such as simulation and in-silico methods, offer promising solutions for accelerating drug discovery.
• 👶 NVIDIA's Clara Discovery suite provides scientists with tools and systems for discovering new drugs.
• 👨‍🔬 Cambridge-1, with its powerful computing infrastructure, aims to support AI and healthcare research collaborations in the UK.
• ❓ Partnerships between pharmaceutical companies, academic institutions, and healthcare organizations are crucial for advancing AI drug discovery.
• 🤩 The use of technologies like genomics, cryo-Electron-microscope-imaging, data analytics, deep learning AI, docking, molecular dynamics, pathology and radiology imaging, and natural language processing are key in the drug discovery process.
• 😫 NVIDIA's DGX SuperPOD, a ready-made AI supercomputer, enables enterprises to quickly set up powerful AI computing infrastructure for their researchers.

Questions & Answers

Q: How can breakthroughs in computer science accelerate drug discovery?

Breakthroughs in computer science, such as simulation and in-silico methods, allow scientists to understand the biological machinery of disease-related proteins and search for potential drug candidates before in-vivo testing. This speeds up the drug discovery process.

Q: What challenges are involved in drug discovery?

Drug discovery faces several challenges. Firstly, finding the specific protein implicated in a disease can be difficult. Secondly, identifying a small molecule that can bind with and activate or deactivate the protein is also challenging. Thirdly, getting the small molecule inside the cell poses obstacles due to natural defense systems. Additionally, understanding how the body absorbs, distributes, metabolizes, and excretes the compound (pharmaco-kinetics) is crucial but complex.

Q: What technologies are used in drug discovery?

Drug discovery utilizes various technologies. Genomics helps select protein targets and determine patient response. Cryo-Electron-microscope-imaging and 3D reconstruction reveal protein structure. Data analytics screens potential chemical compounds. Deep learning AI designs new leads, while docking predicts chemical-protein affinity. Molecular dynamics analyzes protein motion and interactions. Pathology and radiology imaging study biomarkers and drug effectiveness. Natural language processing explores chemical re-purposing possibilities.

Q: What is the purpose of Cambridge-1?

Cambridge-1 is a computing infrastructure being built by NVIDIA to support AI and healthcare research collaborations in the UK. With 400 petaFLOPS of AI performance, it aims to be the fastest system in the UK and a top 30 system globally. It will enable partners, including AstraZeneca, GSK, King's College, NHS, and Oxford Nanopore, to conduct experiments that require significant computing power.

Summary & Key Takeaways

• Drug discovery is a challenging process that takes over a decade and costs billions of dollars, with a high failure rate. Using computer science breakthroughs like simulation and in-silico methods, drug discovery can be accelerated.

• NVIDIA is introducing NVIDIA Clara Discovery, a suite of tools for scientists to discover new drugs. The company also plans to build Cambridge-1, a state-of-the-art computing infrastructure to support AI and healthcare research collaborations in the UK.

• Partnerships with organizations like AstraZeneca, GSK, King's College, NHS, and Oxford Nanopore have been formed to advance AI drug discovery. GSK and NVIDIA are also partnering to establish the world's first dedicated in-house drug discovery AI lab.