|
Abstract:
I hope last week's talk by Yoseph has convinced you that alternative splicing (AS) is a very important piece of cutting edge biological research. In this talk, however, I will try to downplay all the biological subtleties while focusing on the (interesting) machine learning bits of my research. In order to do so, I will first informally introduce the food fair process (a freshly cooked term, but as the name indicates, it has connection to both the Chinese restaurant and Indian buffet processes), and then formulate a closely related generative model of exon abundances for our newly designed AS microarray under the Bayesian framework. I will describe a Markov chain Monte Carlo (MCMC) sampling scheme to attack the problem and show some preliminary results. Both the model and algorithm are still under development so helpful comments and suggestions will be greatly appreciated.