Research Areas

~ Sorin Istrail, Julie Nguyen Brown Professor of Computational and Mathematical and Professor of Computer Science Sciences
Algorithmic Methods in Genomics, Regulatory Genomics, Genetic Basis of Disease, Medical Bioinformatics

My laboratory's main research projects are:
(1) Genomic Regulatory Networks focusing on sea urchin developmental gene regulatory networks, building a high-resolution transcriptome map of the embryo, and inferring logic functions of genomic cis-regulatory code and the principles of information processing of genomic regulation;
(2) Computational Models of SNPs and Haplotypes dealing with HapMap analysis tools design for SNP selection, haplotype phasing, and genome-wide disease associations;
(3) Medical Bioinformatics with focus on comparative immunopeptidomics of humans and their pathogens, genetic determinants of sudden cardiac death and human-rabbit comparative genomics, computational support for pathology diagnosis of cancer.
I am also interested in building a programming language for genomics, designing protein folding algorithms, and continuing John von Neumann's research program towards a new information and computation theory for biological complex systems.

~ Charles Lawrence, Professor of Applied Mathematics
Bayesian Inferences of DNA and RNA Structure and Function

Statistical inference of structure and function for nucleotide polymers is my laboratory’s area of expertise. The lab’s research focuses on nucleic acid structure and function, including studies of the cis acting features of transcription regulation and RNA secondary structure and its relationship to function. We develop and apply Bayesian inference technologies, including exact Bayesian and MCMC sampling algorithms to draw inferences on these systems using genome sequence data. Applications include whole genome inferences of cis-regulatory elements and regulons for single cell species, cis-regulatory modules and regulons in multi-cellular species, cis regulatory features of RNAi mechanisms, and RNA secondary structure landscapes.

~ Benjamin Raphael, Assistant Professor of Computer Science
Computational Cancer Genomics, Biological Sequence Analysis

-Computational Cancer Genomics. Cancer is driven by mutations in DNA including both single letter changes and more drastic, larger scale rearrangements of the genome; e.g. chromosomal inversions, translocations, deletions, and duplications. In some types of cancer, these large-scale rearrangements produce changes in gene structure and regulation that are directly implicated in cancer progression and are targets for cancer therapeutics. In 2003, I began developing algorithms for analyzing genome rearrangements in tumors through a technique called End Sequence Profiling. This work is an ongoing collaboration with Colin Collins, Joe Gray, and Stas Volik at the University of California, San Francisco Cancer Center. In addition to the obvious medical importance, the study of cancer presents numerous challenges that impact nearly every area of bioinformatics.
-Biological Sequence Analysis. The second focus of my research is the development of algorithms for biological sequence analysis, specifically methods for analyzing and annotating DNA and protein sequences including problems in multiple sequence alignment, motif finding, comparative genomics, and structural variation in the human genome.

~ Daniel Weinreich, Assistant Professor of Biology
Evolutionary Genetics and Genomics

I employ population genetics, computational, molecular, microbial and protein biology to explicate the Darwinian paradigm in its most fundamental, mechanistic terms. In particular, I am interested in the consequences on evolution by natural selection arising from functional interactions between mutations within loci or between loci within genomes. Such interactions mean that the phenotypic effect of a mutation vary as a function of what other mutations are already present, a phenomenon called epistasis. My work follows parallel theoretical and experimental lines.
Theoretically I am interested in understanding what forms of epistasis for fitness influence the ability of natural selection to follow mutational and recombinational trajectories to higher fitness genotypes. More abstractly, I am working on methods to characterize the space of all possible patterns of epistasis, with with particular emphasis on consequences for evolution by natural selection.
The absence of mature theory has also left many basic empirical questions unasked and I am exploring some of these in the laboratory, chiefly by measuring the fitness effects of carefully selected sets of mutations singly and in combination. Additionally, in order to understand the mechanistic basis of such functional interactions, I seek to dissect the fitness consequences of individual mutations into a causal chain of effects at the molecular, biochemical, biophysical, cellular and physiological levels. Because evolution by natural selection is an important component of human infectious disease biology, much of my work employs is in human pathogen models, which also offer many technical advantages. Additionally I work in bacteriophage and marine bacteria.
I am actively recruiting students at all levels who possess a background or strong interest in one or more of evolution, population genetics, molecular or protein biology, computational biology or theory.

~ Will Fairbrother, Assistant Professor of Biology
Using Computational Methods to Identify Sequence Elements that are Involved in Gene Expression.

The pursuit of my research goals: exploring mechanisms of signal evolution and understanding the effects of mutation on gene expression will require reliable definitions of gene expression signals. We developed a computational method, RESCUE-ESE, for identifying exonic splicing enhancers in human genes. Although ESE sequence motifs were poorly defined prior to RESCUE-ESE, the literature had provided some insight into how ESEs function. For example, ESEs had been shown to function by compensating for weak (non-consensus) splice sites and this enhancement was lost when the ESE was moved to an intronic location. These observations suggested that ESEs would possess two attributes: They would be A) enriched in exons (relative to introns) and B) enriched in exons flanked by weak (non-consensus) splice sites relative to strong (consensus) splice sites. The RESCUE-ESE method counted word frequencies in the vicinity around splice sites and found 238 hexamers to be significantly enriched in both of these attributes (see A and B above). These 238 hexamers were clustered on the basis of sequence into 10 unique sets.

~ Franco Preparata, An Wang Professor of Computer Science

Following early research in switching and coding, culminating in the discovery of the nonlinear Preparata codes, for the past three decades the focus of Franco Preparata's research has been the design and analysis of algorithms in their most general connotation. With the remarkable evolution of computer technology, his research interests have been correspondingly evolving. He has been deeply interested in fundamental algorithms and data structures, VLSI computation and layout, and parallel algorithms.

Perhaps the most enduring interest has been computational geometry, a spin-off of algorithmic research aimed at the systematic investigation of methods for the most efficient solution of geometric problems. Geometric problems are ubiquitous in human activities. Sporadic, and frequently inefficient, computer solutions had been proposed before, but in the mid-seventies computational geometry emerged as a self-standing discipline targeted at this important area. The goal of computational geometry is to analyze the combinatorial structure of specific problems as the underpinning of efficient algorithms for their solution. The field burgeoned, and in the mid-eighties Prof. Preparata wrote a textbook on the subject that helped establish it in the instructional arena. Today an enormous body of geometric algorithms is known and this knowledge is increasingly indispensable in several applied areas such as geographic information systems, computer graphics, and computer-aided design and manufacturing. Within the last area, Prof. Preparata has also contributed to computational metrology—the assessment of the geometric quality of manufactured parts.

As another example of computer science interacting with other fields, today his main research focus is computational biology (also called 'bioalgorithmics'), an emerging discipline that entails the development and use of mathematical and computer science techniques to solve problems in molecular biology. Since the discovery of the structure of DNA about 50 years ago and the digital underpinning of molecular biology, huge amounts of data have been generated in this field, making it necessary to resort to sophisticated computer science techniques for their analysis.

~ David M. Rand, Professor of Biology
Molecular evolution and ecology; nuclear-mitochondrial coevolution; comparative genomics

We are interested in molecular evolution, comparative genomics and association mapping of continuous traits. Most of our work is empirical, but we employ computational methods using existing software packages (PAML, MEGA, DNAsp), or for specific questions, custom software written in C++ or Perl. A Major interest is how selection acts on mitochondrial genes. We have developed neutrality tests of nucleotide and protein sequence data to infer selection on standing variation. A focus of current research is the coevolution of nuclear and mitochondrial genomes, with particular interest in nuclear genes that function in the mitochondrion and alter fitness effects of mtDNA mutations. We employ laboratory selection experiments to uncover genomic regions that carry allelic variation governing the selected traits. Ongoing work with thermal selection has identified a region of the Drosophila X chromosome that alters fitness at different temperatures; we are dissecting this region with gene mapping methods and analyses of latitudinal clines of haplotype variation.

~ J. William Suggs, Associate Professor of Chemistry and Biochemistry

We are interested in ligand design and synthesis. In particular, we are interested in how DNA molecules can bind in specific ways to organic molecules, especially drugs used in cancer chemotherapy. We also seek to understand how drug molecules can induce specific rearrangements in mammalian chromosomes, since abberant chromosomes are one of the hallmarks of cancer cells. Finally, we have a program underway to make chemically modified DNA polymers to test the limits of molecular information storage. Recent work has resulted in our synthesizing DNA binding drugs whose location in cellular DNA can be determined by electron microscopy.

~ Zhjin Wu, Assistant Professor of Medical Science

Zhijin's primary interest is in developing statistical methods for analyizing data from microarray experiments. She and colleagues proposed the model-based preprocessing method, GCRMA, that adjusts for non-specific background and normalizes raw measurements from microarrays and produces more accurate and precise estimates of gene expression levels. In addition to their original implementation of the method provided through the Bioconductor project, GCRMA has also been incorporated into commercial software for microarray data analysis, such as GeneSpring, and has become a popular alternative to the manufacturer-provided procedure.

Besides gene expression experiments, she continues to work on methods for new applications of the microarray technology, such as detecting single nucleotide polymorphisms (SNPs) and DNA methylation. She is also interested in high-throughput screening (HTS) on cell based assays.

_______________________________________________________ top of page