Geometric deep learning for drug discovery

What if we could design drugs for the 80% of proteins current methods can't target? Geometric deep learning is making it possible.

#1about 2 minutes

Why many common diseases remain untreatable

Over 80% of pathogenic proteins are considered 'undruggable' because traditional inhibition methods are ineffective for them.

#2about 4 minutes

Introducing targeted protein degradation technology

This biotechnology shifts the paradigm from inhibiting symptoms to eliminating the cause of disease by degrading pathogenic proteins.

#3about 7 minutes

Using geometric deep learning for molecular data

Graphs provide a native, non-Euclidean way to represent complex molecular structures and interactions, which is essential for effective machine learning models.

#4about 11 minutes

An overview of the AI drug discovery pipeline

The process involves predicting protein-ligand interactions, generating molecular conformations, and using Bayesian optimization to find optimal candidates.

#5about 1 minute

Optimizing molecular poses with a fitness function

Bayesian optimization and active learning are used to efficiently search the high-dimensional space of molecular rotations and translations to find the best interaction.

#6about 3 minutes

Generating novel molecules with generative models

Generative models create entirely new linker molecules de novo, optimizing for properties like low toxicity and minimal energy during generation.

#7about 3 minutes

Building a cloud architecture for large-scale ML

A custom cloud architecture using AWS spot instances and persistent storage is necessary to handle the immense computational cost of geometric deep learning.

#8about 3 minutes

The key enablers for AI-driven drug discovery

Success requires a combination of high-quality data, advanced algorithms like geometric deep learning, cloud computing power, and an interdisciplinary team of domain experts.

#9about 9 minutes

Q&A on team collaboration and technical choices

The discussion covers the challenges of communication in interdisciplinary teams, the rationale for a multi-cloud strategy, and specific technical questions about the models.