Advances in technology, combined with new discoveries at the molecular level, have significantly changed the nature of biomedical research during the past two decades. Consequently, the pursuit of the understanding of complex disease, its prevention, and treatment, involves addressing complicated questions which span multiple fields of research. Expertise in the realms of clinical practice, molecular biology, chemistry, mathematical modeling, computer science, and statistics, among others, all have roles to play in generating a comprehensive understanding of the processes involved. My role in this research is as a biostatistician and bioinformatician, collaborating with researchers in biochemistry, microbiology, and other biomedical disciplines.
Advances in technology, combined with new discoveries at the molecular level, have significantly changed the nature of biomedical research during the past two decades. New tools give investigators the opportunity to perform experiments that generate many millions of data points. Developments in our understanding of molecular mechanisms of gene expression - for example, in actions of microRNA and in epigenetic modification - have both provided new methods of experimental analysis and added to the complexity of our understanding of the molecular basis of biological function. In pursuing the understanding of complex disease, its prevention, and treatment, we typically address several different questions: Are there genetic signatures that predispose individuals to the disease? Are there environmental factors that alter gene expression in such a way as to increase an individual's risk of disease, and if so, what is the molecular mechanism of this increased risk? Once an individual contracts a disease, what is the mechanism of progression of the disease? The ultimate aims of the research are to use answers to these interrelated questions to develop prognostic tests, interventions, and behavioral modifications to prevent, slow progression of, and treat the disease.
Due to the complexity of these questions, and of the data generated in answering them, it is no longer feasible to achieve these aims by addressing them from the perspective of a single field of research. Consequently, we need teams of collaborative researchers with expertise in their own fields and the ability to communicate and collaborate with those outside of their fields of research.
My role in this research is as a biostatistician and bioinformaticist. As a collaborative researcher, I aim to maintain a full understanding of current statistical techniques, to determine which of those techniques are best applied to problems presented by the generation of data from biomedical research, to maintain the technical skills to perform those analyses, and to present and communicate the results to the research community at large. A key aspect to this collabortive research, is the ability to communicate with researchers outside one's own field; both in terms of understanding the research being performed in the lab, and in terms of being able to explain the choices behind the analyses being used and their results. On occasion, it may be necessary to develop novel statistical techniques in order to perform the analyses required for a particular project.
I am specifically interested in research in complex disease, including cancer and cardiovascular disease. My interest stems both from an "applied" perspective (these are diseases in which our understanding and consequently our abilities to provide treatment have the greatest potential for further progress), and from a "theoretical" perspective (the mathematical relationships between the various causes of the disease and its manifestation are intellectually appealing to me). My particular expertise is in the application of multiple hypothesis testing procedures to high-throughput genomic data. I am interested in the future to combine different sources of genomic data (such as sequence data, gene expression data, microRNA expression data, methylation data, and proteomic data) to achieve a more global understanding of the molecular mechanisms of disease. At the clinical level, this translates to pursuing a unified understanding of the interaction of "traditional" genomics with the molecular effects of environment and its effect on disease.
I provide services ranging from basic data analysis to analysis of high throughput data, database management, and custom scientific programming to biomedical researchers at Marshall University, other academic institutions in West Virginia, and to the wider scientific community.
Data analysis and basic statistical support
Faculty and students in the School of Medicine needing support in analysing their data may contact me for assistance. I strongly encourage you to talk to me during the design phase of your experiment, in order to ensure you collect data and have a plan for data analysis which will accurately test your hypotheses. I also encourage investigators who are in the process of writing grants to discuss the statistical needs of their projects with me during the grant writing phase. For projects that will require extensive statistical support, you should consider including salary support in your budget.
Analysis of microarray and high-throughput data
Genome-wide anlayses such as those performed by microarray and high-throughput sequencing present specific challenges in both management and analysis of the data generated. In particular, the very large number of parallel tests being performed make the interpretation of statistical results for an individual gene, snp, or microRNA highly complex. A number of specialized statistical techniques are available for this scenario. Through the Marshall University Genomics Core Facility I analyze data generated from both microarray and high-throughput sequencing assays performed in the core. This analysis is an integral part of the service provided by the core facility. For data generated from other facilities, including data available in public data repositories such as NCBI GEO. Several investigators with whom I have worked have made significant discoveries using data generated from previous publications and reanalyzed to test their own hypotheses.
Data management and database programming
The various cores and grants with which I am associated support a number of different pieces of computer server hardware, most of which run relational database systems. For large projects which need database support, including custom, web-based interfaces, we can consider developing and supporting those on one of our servers.
Custom scientific programming
For novel or especially complex assays, off-the-shelf software may not provide adequate visualization of data. For these cases, acustom programming in Java and/or R can be provided to produce images suitable for publication or presentation.
I teach BMS 617 ("Statistical Techniques for Biomedical Sciences"). This course runs in the Spring in alternate years.