Group Leaders

Professor, Department of Pharmaceutical Chemistry at UCSF

An overarching interest is to bring chemical reagents to biology, using a combination of computational simulation and experiment. Using a protein-centric approach, we search for new ligands that complement protein structures. Using a ligand-centric approach, we try to find new targets for known drugs. A particular focus is the discovery of reagents to modulate G-Protein Coupled Receptors. Two crazy goals for the next five years are predicting one ligand for every accessible protein target, and for every drug predicting one new target to which it binds.

ANALYSIS OF GENETIC AND NEURONAL CIRCUITS OF FAT AND FEEDING

A major challenge in biology is to understand how gene products form dynamic networks that operate in multiple tissues to coordinate physiology. One such network is the set of genes that determine the balance between energy intake and utilization. Fat regulation involves a complex interplay between feeding centers in the nervous system and the transport, storage and mobilization of fat in the periphery. Defects in energy balance are associated with metabolic disorders, cardiovascular disease, and cancers.

Using mutagenesis and genome-wide RNAi studies, we previously identified over 400 genes that regulate fat levels in C. elegans. These include components of neuroendocrine signaling, channels and transporters, transcription, metabolism and many genes with previously unassigned functions. Our findings highlight the shared ancestry of C. elegans and mammalian fat regulation. More importantly, this work identifies hundreds of new genes that participate in novel fat regulatory pathways. We have now undertaken the daunting task of understanding the modes of function of these fat regulatory genes within the context of intact organisms. These genes affect functions as diverse as metabolism and behavior.

Interests

Pharmacometrics, Quantitative Pharmacology, Nonlinear mixed effects models, Infectious diseases, Multiple sclerosis, Disease progression models

The long-term research goal of the Butte Lab is to solve problems relevant to genomic medicine by developing new methodologies in translational bioinformatics.

1. We have developed bioinformatics methods to take genomic, genetic, phenotypic, and RNAi data from multiple sources and phenotypes and reason over these data. An example of this was in our work in cancer drug discovery published in the Proceedings of the National Academy of Science (2000), on type 2 diabetes mellitus published in the Proceedings of the National Academy of Science (2003), on fat cell formation published in Nature Cell Biology (2005), on obesity in Bioinformatics (2007), and in transplantation published in Proceedings of the National Academy of Science (2009).


 

2. We have developed tools to automatically index and find genomic data sets based on the phenotypic and contextual details of each experiment. Our work on automatic annotations of the data in the NCBI Gene Expression Omnibus was published in Nature Biotechnology (2006), our work in re-mapping microarray data was published in Nature Methods (2007), and our work in deconvolution of whole blood expression was published in Nature Methods (2010).


 

3. We have been developing some novel methods in comparing clinical data from electronic health record systems with gene expression data, as described in Science (2008), using genetic measurements, as described in PLoS Genetics (2010), and using gene-expression data, as described in Molecular Systems Biology (2009), and described in PLoS Computational Biology (2010), and described in the New York Times and International Herald Tribune.

The Helix Group at Stanford is directed by Russ Altman, and focuses on the creation and application of computational tools to solve problems in biology and medicine. Current application projects include the study of structure-function relationships in macromolecular structure, understanding the structure and folding of RNA molecules, and analyzing the relationship of genotype and phenotype, particularly with respect to the response to drugs. Techniques used include knowledge representation, database design, machine learning, natural language processing, physics-based simulation and graph-based modeling/analysis.