Hands-on activity for the XI GEFENOL summer school

Course by Caterina A. M. La Porta and Stefano Zapperi

This hands-on activity shows the basic functionality of GEO2pandas, a python package written by Francesc Font-Clos to retrieve data and metadata from the Gene Expression Omnibus into pandas.DataFrame objects.

The tutorial is part of the XI GEFENOL summer school in Barcelona and is adapted from a tutorial set up by Francesc Font-Clos at the University of Milan.

Before starting

To follow this lecture, you need a modern installation of python, together with jupyter, numpy, matplotlib and some other standard python libraries. The simplest way to install all these packages without interfeering with your current python installation is the Anaconda distribution. Choose python 3.x and your OS, download, install, and you should be good to go.

Once you have install python you should download the entire repository and local install GEO2pandas.

Use a terminal window, cd to the GEO2pandas directory and install the following:

• pip install geoparse == 2.0.1
• pip install pandas == 1.0.3
• pip install numpy == 1.18.1
• pip install -e ./

the last command will install GEO2pandas

Hands-on structure

First part
Starting just with from a keyword of interest (e. g. obesity), I will show how to do some basic exploratory analysis of public transcriptomic data related to our keyword.

1. Searching for datasets: GEO and ArrayExpress
2. Getting the data. Possible issues.
3. Basic analysis 3.1 Differentially expressed genes
4. Basic visualization 4.1 Heatmap of DE genes

Second part
Starting with a keyword of your choice, you will try to reproduce the steps 1-5 above. Problems might arise.