Machine Learning for Regulatory Genomics
This module handbook serves to describe contents, learning outcome, methods and examination type as well as linking to current dates for courses and module examination in the respective sections.
IN2393 is a semester module in English language at Master’s level which is offered in summer semester.
This Module is included in the following catalogues within the study programs in physics.
- Catalogue of non-physics elective courses
|Total workload||Contact hours||Credits (ECTS)|
|180 h||60 h||6 CP|
Content, Learning Outcome and Preconditions
introduce biological mechanisms, experimental assays, and computational models for regulatory genomics. The six lectures are supported with modeling exercises in python. This is followed by (2) an eight-week hands-on project.
The lectures are organized around steps of gene expression:
- Introduction to gene regulation and sequence-based computational models of gene regulation
- Transcriptional regulation
- Chromatin-mediated regulation
- RNA splicing
- RNA modification and degradation
Over these lectures, computational methods are introduced including:
- Fitting procedures of deep neural network
- Convolutional Neural Networks
- LSTM and transformers
- Embeddings for sequence data
- Multi-task learning and transfer learning
- End-to-end learning
- Analytical and visualisation techniques for model interpretation
1. Describe major steps of gene expression from accessing DNA to determining protein abundance.
2. Describe genome-wide assays employed to assess various steps of gene expression
3. Describe the concept of massively parallel reporter assays
4. Describe and apply deep learning methods to perform sequence-based predictions
5. Describe and apply the concept of model interpretation
6. Describe and apply the concept of convolutional neural network
7. Describe and apply the concept of transformers
8. Apply deep learning for sequence-based modeling of a genome-wide assay. Evaluate model performance and provide biological interpretation of its application to real data.
physics, statistics or a related field. One lecture on machine
learning (e.g IN2064; MA4802). Strong interest in
biological and biomedical research questions.
Courses, Learning and Teaching Methods and Literature
Courses and Schedule
|VO||2||Lecture Machine Learning for Regulatory Genomics (IN2393)||Gagneur, J. Heinig, M.||
Tue, 14:00–15:30, virtuell
|UE||2||Exercise Machine Learning for Regulatory Genomics (IN2393)||Gagneur, J. Heinig, M.||
Tue, 15:30–17:00, virtuell
Learning and Teaching Methods
Lectures provide the state-of-the-art of regulatory genomics modeling approaches. These concepts are first applied with in-class tutorials following each lecture. During the project work, these concepts are applied on real biological or biomedical data problems under mentoring of the teaching team. The results of the project work are summarized in a final talk and a written report.
Eraslan et al. Deep learning: New computational modeling techniques for genomics, Nature Reviews Genetics, 2019
Description of exams and course work
- During the project work (motivation, problem solving
capacity, data analysis skills, programming
- At the final presentation (clearness of presentation
and slides, used methods, achieved results). 10 minutes.
- In the written report (conciseness, language, used
methods). 20 pages maximum.
The final mark will be given by the supervisors who attend the final lectures.
The exam may be repeated at the end of the semester.
Current exam dates
Currently TUMonline lists the following exam dates. In addition to the general information above please refer to the current information given during the course.
|Machine Learning for Regulatory Genomics|