We believe that everyone should be able to use generative AI for drug discovery. Yet, researchers struggle to implement this new technology in their daily work.
Learn how generative foundational models can accelerate drug discovery, not in theory, but in a live hands-on demo using the single-cell GPT (scGPT) foundational model.
Learn:
- Where are foundational models useful and becoming crucial to speed up research in biotech and pharma.
- How to quickly get started with scGPT, one of the most promising foundational models in biology today.
- Addressing the technical challenges posed by large models, including model sizes, large datasets, complex dependencies, and GPU requirements.
- How to fetch relevant datasets from the Chan Zuckerberg CELL by GENE
- Discover datasets, keeping the lineage record from the data download to the final result.
- How to extract embeddings from the scGPT model and visualize them in UMAP representations.
- How to perform modern data analysis with VSCode in the cloud with the aid of Github Co-Pilot.
- How to perform downstream machine learning experiments with the aid of MLFlow to extract biological insights using XGBoost on the scGPT embedding layer.
Discover:
- How to set-up in silico gene perturbations and quantify their effect of Crohn's disease across various cell types and organs.
- How the lineage of the obtained result is automatically captured start-to-end in an interactive UI.
- How the results obtained will be reproducible now and any time in the future.
