12-15 October 2026

This course introduces machine learning and deep learning tailored to application within the drug discovery field. Within four days, participants will learn fundamentals of drug discovery progress and applications of AI, machine learning and deep learning within every stage of the drug developmental pipeline. Participants will gain solid theoretical knowledge and practical skills.

The course will cover both machine learning fundamentals and traditional machine learning applications, as well as more advanced models such as graph neural networks. We will cover both how these models work and how to effectively build them using Pytorch. Participants
will learn and practice using -OMICs data with machine learning models, gain hands-on experience on how to use graph neural networks for molecular analysis, gain fundamental understanding of the most popular library for deep learning as well as how to go beyond standard tutorial level datasets and get real-world data. Additionally, participants will get detailed theoretical coverage of the drug discovery field.

PhD students, MSc students, and entry-level scientists who are interested in drug discovery and want to use machine learning within this field. This course would be especially useful for students and scientists who have some knowledge of biostatistics and R/Python and want to broaden their skill set within one of the most exciting and challenging areas. It will be also useful for individuals with technical backgrounds who want to improve their understanding of the drug discovery process and where to get data for their ideas.

- Intermediate experience with Python
- Basic understanding of molecular biology and chemistry
- Basic understanding of data analysis and statistics
- No deep learning experience is required

After this course participants will:

- Understand drug discovery process for small molecules and biological therapies
- Understand ML/DL/AI applications within the drug discovery domain
- Gain theoretical knowledge of machine learning and deep learning
- Use machine learning and explain their predictions for target discovery
- Apply graph neural networks and explain their predictions to understand properties of drug molecules
- Know how to effectively collect drug-discovery-related data and go beyond tutorial datasets

Day 1 - 2-7 PM Berlin time
Introduction to drug discovery process and problem definitions
● The first half of the day will provide theoretical coverage of the drug discovery field and where machine learning and deep learning is used and which kind of models are advancing the field

General intro to machine learning
● Machine learning fundamentals
● Data setup, model design, training, and adequate evaluation
● What can go wrong and how to avoid both common and very tricky mistakes
● Numerical representation of biological data
● Challenges of biological and chemical data

Day 2 - 2-7 PM Berlin time

Deep learning and graph neural networks
● Deep learning fundamentals
● Building blocks and mathematical transformations in deep learning networks
● Overview of model families used in drug discovery
● Introduction to graph theory, message passing and graph neural networks
● Hands-on: using ML for disease classification and disease target identification

Day 3- 2-7 PM Berlin time

GNNs for drug discovery and screening
● Pytorch: its modules and capabilities
● Hands-on: processing molecular data and getting its numerical representation
● Math behind GNNs and expressive power
● Hands-on: setting-up model, model training, data pipeline, evaluation, and prediction explanation

Day 4- 2-7 PM Berlin time

Going beyond tutorial datasets

● This day will provide coverage of different data sources that are used in research and industry for drug discovery
● Hands-on: programmatic access to biological databases and setting up data pipeline