Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Explore the intersection of artificial intelligence and molecular biology in this 53-minute conference talk that examines both the potential and limitations of AI approaches in biological research. Learn how the molecular biology revolution has reshaped scientific understanding of living systems, with explanations now centered on gene and protein identification, and discover how new technologies enable large-scale biological measurements. Examine two key research areas: transformer-based models for understanding genotype-to-phenotype relationships and LLM-based foundational models for cellular identity, including TranscriptFormer trained on single-cell RNA sequencing data. Understand why large language models excel at capturing evolutionary and demographic patterns in DNA sequences but struggle with cellular identity biology, and see how simple parameter-free linear algebra models can outperform billion-parameter systems like TranscriptFormer in downstream cellular identity tasks. Gain insights into combining linear algebra, bifurcation theory, and statistical physics to classify cell fate transitions using single-cell RNA sequencing data, providing a comprehensive view of current challenges and opportunities in computational biology.
