Skip to content

README.md

Latest commit

 

History

History
143 lines (101 loc) · 4.06 KB

README.md

File metadata and controls

143 lines (101 loc) · 4.06 KB
Raw

Setting up the Python track

There are three ways to run the ZebraQ Python lessons. Pick one.

If you are new to programming, use Google Colab. It requires no installation at all and works on any machine with a browser, including a Chromebook or a tablet.

Option 1 — Google Colab (nothing to install)

  1. Open the repository's main README.
  2. Click the Open in Colab badge next to the lesson you want.
  3. Run the cells from top to bottom with Shift + Enter.

The first cell of every notebook installs anything Colab is missing. That takes about a minute for the lessons that use PyDESeq2 (S4 and S5) and no time at all for the others.

Every notebook is self-contained: it downloads the data it needs directly from this repository, and no lesson depends on you having run a previous one. You can open S5 without having opened S4.

Two things to know about Colab:

  • Your changes are not saved back to this repository. Use File → Save a copy in Drive to keep your work.
  • If a session sits idle too long it disconnects and you lose the variables. Just re-run the notebook from the top.

Option 2 — A virtual environment (recommended for local work)

You need Python 3.11 or newer — pydeseq2 0.5.x requires it. Check with python3 --version.

git clone https://github.com/enriquea/ZebraQ.git
cd ZebraQ

python3 -m venv .venv
source .venv/bin/activate          # Windows: .venv\Scripts\activate

pip install -r setup/python/requirements.txt

jupyter lab

Then open any notebook in lessons-py/.

To leave the environment later, run deactivate.

Option 3 — conda

git clone https://github.com/enriquea/ZebraQ.git
cd ZebraQ

conda env create -f setup/python/environment.yml
conda activate zebraq-py

jupyter lab

Checking it worked

Run this in a Python shell or a notebook cell:

import pandas, numpy, matplotlib, seaborn, scipy, sklearn, pydeseq2
print("pandas  ", pandas.__version__)
print("numpy   ", numpy.__version__)
print("pydeseq2", pydeseq2.__version__)

If that runs without an error, you are ready.

Troubleshooting

ModuleNotFoundError even though pip said it installed. Your notebook is probably using a different Python than the one you installed into. Check with:

import sys; print(sys.executable)

If that is not the Python inside your .venv, register the environment as a kernel and select it in Jupyter:

python -m ipykernel install --user --name zebraq-py --display-name "ZebraQ (Python)"

PyDESeq2 fails to install with No matching distribution found for pydeseq2==0.5.4. You are on Python 3.10 or older. PyDESeq2 0.5.x requires Python 3.11+, and because requirements.txt pins the exact version, pip has no older candidate to fall back to — it fails outright rather than silently installing something incompatible. Create the environment with a newer interpreter:

python3.11 -m venv .venv          # or any 3.11+ you have installed

The conda option already pins python=3.11, so it is unaffected.

A notebook cannot find the data files. Run Jupyter from the repository root, and open notebooks from lessons-py/. The notebooks look for ../data/, and fall back to downloading from GitHub if that path does not exist — so an unexpected download usually means your working directory is wrong.

Plots do not appear. Make sure you are in Jupyter and not a plain Python shell. If a figure still does not render, add %matplotlib inline at the top of the notebook.

Which package replaces which R package

R / Bioconductor Python
base vectors, lists list, dict, tuple, set
matrix, array numpy
data.frame, dplyr, tidyr pandas
factor pandas.Categorical
ggplot2 matplotlib + seaborn
ggpubr scipy.stats + manual annotation
readxl, openxlsx pandas + openpyxl
microbenchmark %timeit
DESeq2 PyDESeq2
EnhancedVolcano matplotlib + adjustText
pheatmap seaborn.clustermap
sessionInfo() session_info.show()