Prerequisites

What this course expects

This course trains a single skill: taking someone else’s biological problem, interviewing them until the question is precise, then directing and verifying an AI agent that does the analysis for you. The scarce ability is specification, direction, and verification — not writing code by hand. So the prerequisites are not a programming test. What we ask you to bring:

• Direct and check an AI agent. Be ready to give an AI assistant a clear, specific instruction, read what it produces, and judge whether the result is sensible. No prior experience with AI coding agents is required — Module 1 gets everyone started.
• Read and interpret code and results at a basic level. You will rarely write code from scratch, but you must be able to read it and sanity-check what an agent gives you: follow the logic of a short script, spot an obvious mistake, and distrust a number that looks too clean. You do not need to produce Python fluently — you do need to be able to follow it.
• Scientific and biological curiosity. Enough interest in the biology to want the answer to be right, and enough scepticism to check it rather than trust it.
• Willingness to work on someone else’s problem. A hard rule of the course: you may not use your own data or research problem. Every lab and the final project use a problem that belongs to someone else, which you reach by interviewing them. Come prepared to interview a real data owner and to state a context that is not already in your own head.

On programming: Python is not a hard entry requirement this year. The examined skill is directing an agent, not coding by hand. Even so, comfort reading code and outputs makes the labs far easier — if you have never seen Python, the optional refresher below gets you to the “I can read and check this” level, which is all you need.

Be prepared

We recommend becoming familiar with the following before the course starts:

• Browse the SciLifeLab Data-Driven Life Science (DDLS) initiative to understand national priorities and the concept of the data life cycle, which is central in this course.
• Get an AI coding agent running on your own machine (e.g. Claude Code, OpenAI Codex, Gemini CLI, or Cursor) and make it do one small real task. Installing it and trying it once is the single best preparation for the labs.
• (Optional) Refresh your ability to read Python — variables, control flow, functions, and how a short data-analysis script is put together — so you can follow and check what an agent writes. See the resources below.

Technical setup for labs (all online):

• A computer you are allowed to install software on (you will run an AI coding agent locally)
• Reliable internet access and a modern browser (e.g. Chrome)
• A Google account (for Google Colab and Drive storage)
• An account with at least one AI assistant / coding agent (e.g. ChatGPT, Claude, or Gemini); free tiers are enough to start
• A GitHub account for versioning and sharing notebooks/code

Optional: getting comfortable reading Python

You will not be asked to write much code, but being able to read it — and notice when something is off — makes every lab easier. If Python is new or rusty, use any of the options below to reach that level. There is no submission requirement; this is for your own preparation.

Option A: Guided practice with AI

Use curated prompts to accelerate review: Learn Python with ChatGPT. Treat AI output critically—run code, fix errors, and keep notes of what you clarified. This is itself good practice for the course: you are directing and checking an AI rather than memorising syntax.

A comfortable-to-read checklist — you should be able to follow code that uses these, not necessarily write it from memory:

• Running cells in Colab / Jupyter
• Using print, f-strings, and basic input/output
• Lists, tuples, dictionaries, sets (creation, indexing, iteration)
• Control flow (if, for, while, list comprehensions)
• Defining functions; understanding scope and return values
• Importing standard libraries (math, random, json, pathlib)
• Basic plotting with matplotlib or seaborn
• Reading CSV/TSV data with pandas
• Simple error handling (try/except)

Option B: Traditional refresher

Use the interactive notebook below or other beginner tutorials (e.g. Python docs, Software Carpentry). Progress until the checklist above feels easy.

Interactive Notebook

Quick Python basics:

Or load locally in Jupyter Lite: + Show Notebook in Fullscreen Mode

If you are new to programming, you may also watch this introductory video:

Quick Quiz (self-check)

x = 5
print(f"Value: {x}")

squares = [i*i for i in range(5)]  # [0,1,4,9,16]

def mean_or_none(values):
    if not values:
        return None
    return sum(values)/len(values)

import matplotlib.pyplot as plt
x = list(range(5))
y = [i**2 for i in x]
plt.plot(x, y)
plt.xlabel('x')
plt.ylabel('x^2')
plt.show()

import pandas as pd
df = pd.read_csv('data.csv')
print(df.head(3))

try:
    n = int(user_input)
except ValueError:
    n = None