This repository contains some simple analysis code that can be run on Human Brain Atlas data.
This is a python example that computes the gene expression correlation between all samples from a single donor. There is also a script for generating a heatmap of the matrix. It specifically runs on the contents of any of the donor zip files found here: http://human.brain-map.org/static/download/.
To run this example, first unzip the data for one of the donors. Then run the following:
$ python compute_sample_correlation.py -p Probes.csv MicroarrayExpression.csv correlation.npz
$ python make_heatmap.py -o Ontology.csv -s SampleAnnot.csv correlation.npz heatmap.png
Inside the python directory you'll see find_best_probes.py. This is a function used by compute_sample_correlation.py to pick one probe per gene to be used during correlation computation. You can run it directly as follows:
$ python find_best_probes.py Probes.csv MicroarrayExpression.csv best_probes.csv
All of the python code has been tested with python 2.7.3 and requires the following libraries to be installed:
- numpy
- scipy
- matplotlib