

Breaking Down AI’s Role in Genomics and Polygenic Risk Prediction - with Dan Elton of the National Human Genome Research Institute
9 snips Apr 1, 2025
Dan Elton, a Staff Scientist at the National Human Genome Research Institute, dives into the fascinating interplay between AI and genomics. He discusses AlphaFold, which accurately predicts protein structures, revolutionizing drug development. Elton also explores how machine learning enhances polygenic risk prediction, allowing for improved disease risk assessment based on genetic markers. Additionally, he touches on AI's role in gene editing and the data challenges faced in managing sensitive genetic information, paving the way for future advances in personalized healthcare.
AI Snips
Chapters
Transcript
Episode notes
AlphaFold Revolutionizes Protein Engineering
- AlphaFold, from DeepMind, revolutionizes protein structure prediction, though not directly analyzing genetic sequences.
- Its protein folding modeling transforms drug development and protein engineering, connecting to genetics through encoded amino acid sequences.
Linear Models vs. Neural Networks in Genetics
- Linear models have been effective in analyzing genetic sequences for disease risk and trait prediction.
- Neural networks may offer more precise predictions, but data availability and computing power pose limitations.
AI for Biosecurity
- MIT's Ethan Alley developed AI to detect engineered DNA sequences.
- This has implications for biosecurity, such as determining a virus's origin (lab-made or natural).