This repository contains code and supplementary materials for paper named "Gene co-expression in breast cancer: a matter of distance". Alfredo Gonzalez, Jose Maria Zamora-Fuentes, Enrique Hernandez-Lemus, Jesus Espinal-Enriquez
Install miniconda:
$ bash sh/install-miniconda.sh
Install julia:
$ bash sh/install-julia.sh
We shared expression matrices as public data in:
http://espinal3.inmegen.gob.mx:5270/genecorr/subtipos-originales/
to save locally on Data/Expression:
$ bash sh/download-data.sh
$ bash sh/01-cleandata.sh
Outputs are saved on:
$ ls Data/clean/
This step can be slow.
$ bash sh/02-clustering.sh
Execute in backgorund to best performance:
$ bash sh/02-clustering.sh &> salida.log &
Clustering Outputs are saved on:
$ ls Data/Clustered
Clustering by independent chromosome was named with chr label, and clustering by all chromosomes was named as all label in each subtype directory.
Pearson triangular matrices are in list format are saved on:
$ ls Data/Pearson/
Optional: If you don't have computing hardware. You can download our results with
$ bash sh/download-pearson.sh
Warnning: Pearson correlations can use 30GB of space.
To get heatmaps of pearson corelation for every subtype:
$ bash sh/03-plot-corr.sh
Plots are saved on:
$ ls Plots/Healthy/ # for example
$ python py/04-plot-clusters.py
Plots are saved by chromosome as:
$ ls Plots/chr1/Healthy-chr1-rainbow.png # for example
$ ls Plots/chr1/Basal-chr1-rainbow.png # for example
$ python py/05-plot-gstart.py
Plots are saved by chromosome as:
$ ls Plots/chr1/Healthy-chr1-gstart-cum.png
$ ls Plots/chr1/Healthy-chr1-gstart-heat.png
Files with cum suffix are cummilative plots of elements in every cluster. Moreover heat suffix are heatmap plots by cluster. The colors are the same for each cluster in cum and heat.
$ python py/06-plot-scatt.py
Plots are saved by chromosome as shown:
$ ls Plots/chr1/Healthy-chr1-pdist.png