Gene Expression Prediction

 We aimed at predicting Gene Expression based on epigenetic changes to Histone proteins present in the DNA. A 1D convolutional neural network model was built to take histone modifications as input and to predict whether the corresponding genes would express or not. The motivation behind the project was these epigenetic changes to histones can be caused by various external environmental factors like smoking or pollution. If we could quantify which changes actually resulted in modified expression we could look into ways to develop methods to stop those changes. This could greatly help in drug discovery for cancers and other diseases.

More details about the project can be found in the paper here. You can also look at the presentation below which was presented in IEEE Symposium on Deep Learning to describe our work.

This work was done as part of my junior year research internship at the Machine Learning Lab, Department of Electrical Engineering, Indian Institute of Technology Roorkee. I would like to thank Dr GN Pillai and Dr Maya S Nair for hosting me and providing guidance throughout the process.