The course provides both theoretical and practical foundations for the effective use of ChatGPT and other Large Language Models (LLMs) in scientific research, bioinformatics, and medical education.
Through thematic modules and hands-on activities, students will learn to:
-
understand how LLMs work;
-
design effective prompts;
-
use ChatGPT to generate, debug, and document code;
-
integrate LLMs into reproducible research workflows, from single-cell analysis to the creation of GitHub repositories.
The course is organized into five modules featuring video lectures, interactive examples, and three hands-on exercises.
Day 1 – Module 1: Understanding Large Language Models
What are LLMs and how do they work. Differences between general-purpose models (ChatGPT, Claude, Gemini) and code-oriented ones (CodeLlama, StarCoder). Reasoning limitations and potential applications in scientific and bioinformatics contexts.
Day 2 – Module 2: Prompt Engineering for Scientific Coding
How to structure effective prompts. The importance of context (OS, programming language, data format). Prompt examples for Bash, Python, and R in biomedical environments. Strategies to reduce hallucinations and obtain executable, reproducible code.
Day 3 – Module 3: Debugging with ChatGPT
Understanding different types of bugs (syntactic, logical, and silent). Using ChatGPT to identify and fix errors. Techniques to provide context (error traceback, print statements). Creating test datasets for code quality control.
Day 4 – Module 4: Advanced Use Cases and Documentation
Reverse engineering code with ChatGPT. Automated generation of README.md files and pipeline documentation. Using ChatGPT to explore libraries, tools, and packages in computational biology. Distinguishing between exploratory research (ChatGPT) and verified documentation (Google, PubMed).
Day 5 – Module 5: Practical Applications (Hands-on Exercises)
Exercise 1: Using ChatGPT to create a JupyterLab framework in Docker.
Day 6 – Module 5: Practical Applications (Hands-on Exercises)
Exercise 2: Supporting single-cell RNA-seq analysis in Drosophila with ChatGPT.
Day 7 – Module 5: Practical Applications (Hands-on Exercises)
Exercise 3: Creating an academic website (e.g., a research lab site) via GitHub Pages generated with ChatGPT.
At the end of the course, students will be able to:
- Describe how LLMs and ChatGPT work
- Design effective prompts for code generation
- Use ChatGPT to generate, debug, and document code
- Identify logical and syntactic errors through ChatGPT-assisted debugging
Multiple-choice quizzes at the end of each module.
Hands-on exercises in code generation and testing.
Guided analysis of real errors (debugging with ChatGPT).
Basic knowledge of Python or R.
Familiarity with bioinformatics analysis tools.
Basic understanding of Docker and GitHub is recommended but not required.
Alessandri’ Luca
Per domande o problematiche tecniche relative al corso e alla piattaforma contattare:
edvancedeh@unito.it
Contatto del docente del MOOC: