In this tutorial, we'll delve into the depths of biological data analysis. Using publicly available datasets, we'll use machine learning to try to solve one of life's biggest mysteries: that of completing the wiring diagrams of genetic regulatory networks.
Genes that fire together wire together
In every living cell, there are genetic regulatory networks that dictate how genes are turned on and off. This networks have evolved to help the cell to fine-tune the number and speed of the biomolecules that make up the cell. Despite studying gene networks for more than 30 years in model organisms, the community still faces some problems. The problem we're going to address in this tutorial is to try to make guesses of the "missing wires" of this gene networks.
We'll be using the Keras API to build our neural nets and pandas / numpy / sci-kit learn to wrangle through this massive datasets. Using publicly available RNAseq datasets, we'll train a neural network to predict the biological module of some of the missing nodes in the network. We'll also use the NetworkX library to work with the genetic networks.