EP 62

Predicting Toxicity

This episode explores the challenges of predicting toxicity in drugs, particularly when it comes to rare liver diseases. Dr. Jake Chen's research highlights the limitations of current AI models and proposes a new framework for predicting toxicity. The discussion delves into the complexities of human biology and the need for a more nuanced approach to drug development.

toxicity predictionAI modelsliver metabolismreactive metabolitespersonalized medicine

play_arrowListen to episode

Key Concepts

Toxicity Prediction

Current AI models are limited in their ability to predict toxicity
The liver's metabolic processes can lead to the creation of reactive metabolites that can cause toxicity

Dr. Jake Chen's Research

Proposes a new framework for predicting toxicity that takes into account the liver's metabolic processes
Involves a multilayered approach that includes predicting the metabolites, mapping their transport, and analyzing the biological response of lab-grown organoids

Agentic AI and Multimodal Models

Can help to improve the accuracy of toxicity predictions
Provide a more comprehensive understanding of the complex interactions between drugs and the human body

Data Sources

RNA sequencing and cell painting can provide new insights into the interactions between drugs and the human body
Can help to overcome the limitations of current models and improve the accuracy of toxicity predictions

Personalized Medicine

The ultimate goal of this research is to create a system that can perfectly match the molecular puzzle piece to the individual patient
Can help to eliminate adverse reactions and redefine what it means for a drug to be safe

Episode Summary

check_circleCurrent AI models are limited in their ability to predict toxicity due to their focus on the parent compound and lack of consideration for the liver's metabolic processes.
check_circleThe liver's role in metabolizing drugs can lead to the creation of reactive metabolites that can cause toxicity, which is not accounted for in current AI models.
check_circleDr. Chen's research proposes a new framework for predicting toxicity that takes into account the liver's metabolic processes and the creation of reactive metabolites.
check_circleThe new framework involves a multilayered approach that includes predicting the metabolites, mapping their transport, and analyzing the biological response of lab-grown organoids.
check_circleThe use of agentic AI and multimodal models can help to improve the accuracy of toxicity predictions and provide a more comprehensive understanding of the complex interactions between drugs and the human body.
check_circleThe development of more advanced AI models and the integration of new data sources, such as RNA sequencing and cell painting, can help to overcome the limitations of current models and improve the accuracy of toxicity predictions.
check_circleThe ultimate goal of this research is to create a system that can perfectly match the molecular puzzle piece to the individual patient, eliminating adverse reactions and redefining what it means for a drug to be safe.

Full Transcript

Discussion

Join the discussion — sign in to leave a comment.