Domestication as a model for evolution.
Development of bioinformatics applications for analysis of large-scale genomic data.
As a consequence of these interests, by laboratory encompasses both "wet work" along with development of bioinformatics applications.See Internet Links for links to some of the bioinformatics resources.
For agriculture, the chicken has undergone human directed selection affecting two major traits: egg laying and meat production. This work is being done in collaboration with Drs. Sue Lamont, Mike Persia and Max Rothschild at Iowa State University and Chris Ashwell at North Carolina State University. We are examining responses to heat stress in a variety of chicken lines, including the Heritage and Ross708 birds already described, along with lines maintained at Iowa State University including an the Egyptian Fayoumi line, a Broiler line and a Broiler x Fayoumi advanced intercross line. We will also be examining the response of a commercial egg laying line to heat stress. The objective of this work is to define heat stress responsive genes in the various chicken lines, map quantitative trait loci in response to heat stress and explore epigenetic modifications arising from heat. We anticipate that this work will define alleles that could reduce mortality and production losses due to the anticipated increase in the severity and length of heat waves. In addition, the control populations will allow us to continue our comparative studies of the differences between modern and heritage type birds.
Another part of this effort is to genotype chickens raised in backyard flocks in equatorial Africa. These birds have been under very different selective pressures compared to those selected for large scale Western production. By sampling birds across different environments, we anticipate identifying genetic variants that may have been selected on the basis of a variety of selective pressures including temperature, water availability, feed and disease. The intent is to help African farmers select for chickens that are better adapted to their environments.
This work has largely grown out of the need to analyze very large amounts of expression data. For example, our current RNA-seq transcriptome studies have generated over 1 billion sequences that must be analyzed to make sense out of the data. My efforts have largely focused on means to analyze such data in the context of biochemical and signaling pathways, along with developing new ways to use text-mining to aid the discovery process.
Birdbase is a collaborative effort between researchers at the University of Arizona and the University of Delaware to provide online resources for genetic and genomic studies of birds. Currently, Birdbase provides gene pages with information about each gene, a clearing-house for gene names (Nomenclature committee), an atlas of gene expression patterns (all of our expression data is available through this resource) and a manually curated database of gene ontology terms. In addition, the impact of Birdbase is reaching beyond birds to other Archosaurs, including the Crocodilians.
Pathway Analysis: As high throughput technologies have provided vast amounts of genomic and transcriptome data, it has become important to provide tools that help biologists understand how to integrate this data at a biological level. One means for integration is placing gene expression patterns in the context of metabolic and signaling pathways. By understanding all of the metabolic and signal transduction pathways active in cells, tissues and organs we can begin to explain biological phenomena at a more comprehensive level. Achieving this will require improved bioinformatics resources. This work provided the basis for our current funding from the National Science Foundation that aims at improving visualization tools for pathway analysis. The objective is to use Bubbles, an existing set of tools for streamlining and visualizing software development, to visualize pathways. The advantage of the Bubbles package is it will allow users to interact with dynamic web pages where the symbols representing pathways and gene products are actively connected to underlying information. This will allow users to map expression data to pathways, and visualize how changes in gene products will impact the flow of metabolites or signals through the pathways. In addition, the system will be designed to allow users to visualize changes across pathways instead of limiting visual analysis to individual pathways (as do most current systems). This should greatly facilitate a biologist ’s ability to make inferences and develop new hypotheses based on expression data.
eGIFT and Text-Mining: Massive amounts of data are warehoused in journal articles. The objective of the eGIFT project, that is the result of an active collaboration between Drs. Vijay Shanker, Oana Tudor and myself, is to provide a gene centric web interface that helps biologists rapidly understand the functions of genes of interest (http://biotm.cis.udel.edu/eGIFT/index.php). To convey such information, we developed the concept of the "informative Term" (iTerm). iTerms are single or multiple word(s) that are found at high frequency in the abstracts about a target gene, when compared with the background of words found in all abstracts about every gene. We anticipate that other biologists will find this a useful approach for understanding the large gene lists generated by current methods.