Dates: 15 – 26 August 2022.
Location: Trondheim, Realfagbygget (NTNU).
Organiser: Martin Kuiper.
Evaluation: Each student will submit an individual report, in the form of an executable notebook, on which they will be assessed.
Course webpage: HERE.
Course code: BI8040.
Registration: Closed.
Registration deadline: 8 July.
Credits: 7,5 ECTS.
Lecturers:
- Denis Thieffry (ENS, Paris)
- Martin Kuiper (NTNU)
- Anna Niarakis (Université Paris-Saclay)
- Vincent Noël (Institut Curie)
- Laurence Calzone (Institut Curie)
- Pablo Rodríguez Mier (U Heidelberg)
- Aurélien Naldi (INRIA)
- Åsmund Flobak (NTNU)
- Astrid Lægreid (NTNU)
- Rune Nydal (NTNU)
- Eirini Tsirvouli (NTNU)
- John Zobolas (UIO)
Course description
This course will introduce students to the principles of Boolean mathematical modelling, using a variety of software tools and workflows to build, train and execute logical models for in silico experiments. Logical models of regulatory networks of cell lines or tissue samples can reliably represent these systems, as judged by their stable states and responses to perturbations, for instance simple or combinatorial drug treatments. The process to build such models, optimise their performance and analyse their behaviour involves a variety of tools and approaches that students will learn in a project-based manner. Central to the work is the use of an electronic notebook (Jupyter) in combination with a virtual computing environment (Docker image), which streamlines the course work, warrants reproducibility of results, and forms the basis for individual student reports. The course draws on approaches and technologies that are currently used and developed in the NTNU DrugLogics project (www.druglogics.eu). This research combines knowledge management, Boolean modelling, experimental testing and assessment of relevance for pre-clinical (biotechnological) drug development and clinical cancer diagnosis and treatment. The course will exemplify how such approaches can be used in both the biotechnological and biomedical sectors, such as preclinical drug discovery and repurposing; clinical development of diagnosis and (combinatorial) treatment of cancer and inflammation-based diseases; and the development of personalised patient models (‘digital twins’). The course will also address the consequences of knowledge transformations: the shift from knowledge in scientific papers (hidden) to knowledge bases (knowledge fragments) and actionable Prior Knowledge Networks (logical models). Students will be challenged to consider societal implications of logical model based predictions from Responsible Research and Innovation perspectives.
Course program
The course will have an on-site part (August 15 to August 26) and two off-site parts (the week prior to the course (mandatory reading of papers) and 10 working days after the course (production of a personal eNotebook report).
Day 1: Introductory lectures, introduction to team-based learning-sessions
Day 2-5: Combination of lectures/TBL and supervised student group work with tools and resources
Day 6-7: Project-work: develop and characterise logical models, and their implications for knowledge discovery
Day 8: Introduction to RRI aspects of the work
Day 9: Project-work: develop and characterise logical models, and their implications for knowledge discovery
Day 10: Presentations.
The students will use an eNotebook for the project and reporting. Students write their individual reports after the course finishes, deadline for submitting the report: 7 September.
Learning outcomes and competence
The course content will focus on theoretical principles as well as existing tools (GINsim, MaBoSS, etc.) and resources (CoLoMoTo toolbox) that are currently available for logical model simulations. An overview of resources and tools for knowledge management will be presented, to allow students to build and/or extend a comprehensive logical model. Several more advanced logical modelling and analysis tools (The NTNU logical modelling software pipeline; BioLQM software suite – MaBoSS, UPMaBoSS, Pint) will be applied in projects that students can design themselves or choose from a project portfolio. Logical model simulations will also make use of experimental data (RNAseq, proteomics, …) that can be used to inform the model and optimise its logical rules and topology to reliably represent the corresponding biological system. The project work will exemplify how computational reasoning can be used for (large scale) hypothesis management by using logical modelling (large scale) hypothesis management for interpretation of biotechnology-/biomedicine experimental data and for design of new experiments.
Students will know how to use software tools like GINsim (http://ginsim.org) for building logical models and simulating logical model behaviour under various experimental conditions. They will know which public resources exist to find information from which to build logical models, and they will know how to use scientific literature to further extend logical models. Students will understand for what purpose logical models can be used, and how logical models can help them in their research. They will be able to apply this knowledge in the design of model-based experimental research. Students will understand Responsible Research and Innovation dimensions of model-based Systems Biology and Systems Medicine approaches, with respect to using and disseminating research results and user-perspectives of trust and confidence in these results.
Prerequisites
MSc either in biotechnology, biology, systems biology, biomedicine, or computational biology.
Participants should have some basic knowledge of command lines (unix/linux) and python programming; this basic knowledge can be gathered by following online tutorials:
https://swcarpentry.github.io/shell-novice/
https://datacarpentry.org/python-ecology-lesson/ ([lessons 1-6 and lesson 9])
Participants should be able to install and use software tools for data analysis.
They should also have basic knowledge about bioinformatics-based data analysis, and statistical principles.
An experience with R/Bioconductor is useful in this respect.
Although this is a PhD-level course, Master students can also complete it.