Gene Expression Prediction from Genotype

You can generate predictions of gene expression on your own machine using our MATLAB code.

NOTE:

THIS SOFTWARE IS FREELY AVAILABLE FOR ACADEMIC USE ONLY, and is distributed under the The GNU Lesser General Public License.

Requirements

1. Matlab 2010

Download Instructions

1. Download the compressed file
2. Unzip and extract its content (e.g. tar xvfz GenoExp_Only_Code.tgz).

Contents

The compressed directory contains the following sub-directoris:

• Covariates - containing all the SNP covariate matrices for all genes in our paper
• Response - containing all the centered expression profiles for all genes in our paper
• Metafeatures - containing all the SNP functional metafeature matrices for all genes in the paper
• Models - containing all the pre-trained predictive models for all genes in the paper
• Code - containing all the MATLAB code necessary to either:
• Train your own models
• Predict expression given one of the pre-trained models

Usage

The code can be used in two ways:

• Training mode - the user inputs SNP covariate and expression data and the output is a trained predictive model that can be used to predict expression data of unseen individulas using only their genotyoe data
• Predict mode - the user inputs the SNP covariate data and a trained model (either user trained or pre-trained) and the output is a prediction of expression profile

---

For troubleshooting and FAQ please visit our executables FAQ page.