Contact Name: Samuel Danziger
Contact Email: sam.danziger[at]seattlebiomed.org
Department: Seattle Biomed
We are exploring exon arrays on serum response cells to detect alternative splicing events that act as disease biomarkers and possibly give clues to the mechanism of disease progression. We are currently applying this technology to investigate ALS (i.e. Lou-Gehrig Disease) in mice and Alzheimer’s Disease in humans. We expect these techniques to transition from exon arrays to RNA-seq as the technology progresses. The idea is to train a discriminating classifier to predict ALS on the basis of features derived from the exon array. We have currently used the FIRMA algorithm to predict alternative splicing events and an ensemble of classifiers to make predictions. Techniques developed in this project are also applied to HIV related data sets.
10-week Project #1: For this project, the student would expand a deep learning module written in [R] (or other language) with an eye to using the hidden layer(s) to improve the classifier and possibly identify gene network pathways. There would be plenty of opportunity to apply / develop / expand machine learning and analysis techniques for both disease classification and biomarker discovery.
Experience with [R] statistical programming language.