Example Use Case of AutoML Assistant

Please refer to the Usage and Limitations section before using AutoML assistant. A license is required to use AutoML Assistant.

Identify biomarkers and predict cancer drug response by analyzing gene expression across treatment cohorts

In this use case, you’ll learn how to use the AutoML Assistant to analyze gene expression data from TCGA cohorts to uncover biomarkers associated with treatment response. The Assistant guides you through the end-to-end workflow—from data engineering to model training and evaluation—helping you identify key genes linked to therapeutic outcomes and generate predictive models that can inform precision oncology research.

All you need to do is provide a simple prompt with your request and/or instruction. You can:

• Mention TCGA cohorts with “@”

• Describe research problems using natural language. Below are some examples:

• “Please describe the @5FU_responder and @5FU_non-responder cohort”

• “Please use the two cohorts @5FU_responder and @5FU_non-responder to build a predictive model that classifies whether a cancer patient will respond to 5FU treatment”

• “Please compare expression of TBL1XR1 gene between @5FU_responder and @5FU_non-responder using basic statistics”

AutoML Assistant would respond with:

• Overview of the experimental design for the drug response prediction

• Model performance metrics

• Drug response biomarkers inference

• ML workflow details