| Angel Alvarez: Bayesian networks, algorithm development, transcriptional regulatory networks. Angel is developing an analysis platform called ACORN which can automatically detect connections between transcription factor activity and gene expression. The key challenge here is that we never observe the transcription factor’s activity, so Angel is developing novel inference mechanisms to overcome this problem.
|
| Greg Boggy: Microfluidics, multiplexed assay design, dynamic Bayesian networks. Greg is creating an intelligent experimental system that will not only run the experiments such as cell culture and quantitative PCR, but also model the results and design new experiments based on the observed data.
|
| Andrew Hodges: Bayesian networks, user interface development, gene expression analysis. Andrew is developing a Bayesian network tool called MARIMBA that is able to automatically and systematically trim and augment known gene regulatory pathways given gene expression data. He is also demonstrating the connections between known signaling pathways in databases such as KEGG with data derived regulatory networks. Andrew is coadvised by Yongquin (Oliver) He at the UM department of microbiology.
|
| Raghu Kainkaryam: multiplexed experimental design, coding theory, dynamical systems, particular emphasis on drug discovery and plant genetics. Raghu has developed a novel algorithm named QUAPO that uses biophysical models of drug/target interactions to decode efficient and robust multiplexed drug screens.
|
| Weijun Luo: information theory, gene expression analysis with particular depth in MYC and BMP signaling. Weijun is inventor of the mutual information algorithm MI3, which is currently the top scoring algorithm for identifying regulatory co-factors from gene expression data.
|
| Abhik Shah: Bayesian networks, active learning, and intelligent network prior analysis. Abhik, frustrated with the lack of flexibility and function of the current generation of Bayesian network learning engines, has created a new network learning platform called PEBL. This platform forms the basis for an intelligent user interface called BUBBLE, that not only learns from the data but derives intelligent questions to ask the user when it needs help.
|
| Ryan Welch: Bayesian networks, algorithm development, biomarker identification. Ryan has developed a blazingly fast topologically constrained Bayesian network biomarker identifier called Racecar or MarkIt. This biomarker tool can scan datasets such as SNP data or gene expression data containing hundreds of thousands of variables to quantitatively score biomarkers. Ryan is coadvised by much of the University of Michigan.
|
| Peter Woolf: Bayesian networks, control theory, algorithm development, stochastic simulations. Peter is developing novel visualization environments for analyzing and interpreting the predictions from Bayesian networks to make them more useful to a general audience. Peter is also a compulsive organizer and writer.
|
