Biostatistics and Informatics
The goal of the Biostatistics and Informatics Shared Resource is to support the diverse statistical, informatics and bioinformatics needs of researchers at the University of Hawaii Cancer Center. The Resource was established with the recent merger of the Biostatistics and Informatics Shared Resources. They were consolidated to offer more seamless service on integration of data across sources, including high-dimensional data and in the design and analysis of projects. The co-directors manage the Resource, with Dr. Wilkens providing expertise in Biostatistics and Dr. Okimoto in Bioinformatics.
The Resource provides consultation in the following areas:
- study design
- data collection
- data management
- data analysis
- statistical programming
- bioinformatics and systems biologyinterpretation and dissemination of results
- clinical trials protocol support
The Resource provides expertise for the storage and management of data of all types, which includes customized databases for the inventory and tracking of biospecimens for large population-based cohorts, and the management and tracking of clinical trials for Clinical Sciences and Translational Research Services. Members also provide consultation on data analysis from complex study designs, such as longitudinal, multi-level and clustered studies.
The Resource offers specialized expertise in cancer bioinformatics and systems biology for the analysis and modeling of the high-throughput data sets generated by genome-wide profiling experiments. In terms of infrastructure, the Resource has led in the development of the Cloud Computing for Cancer Research (3CR) scientific computing network, which supports the growing need for accessible, high-performance computing. The considerable overlap between biostatistics, bioinformatics and informatics and the resulting synergy between these functions will deliver a wide-range of quality services to researchers in a timely and cost-effective manner.
Statisticians and bioinformaticians of the Resource engage in the research and development of computational methods that are important to ongoing and future research at the Center. An area of interest for Dr. Wilkens is how to correct for bias in model parameter estimates when the exposure variable is measured with error. The role of dietary intake and physical activity in cancer etiology is of interest to a number of Center researchers, and these variables are measured with error. Dr. Okimotoâ€™s interests lie in development of signal processing algorithms that enhance the identification of genes in large genomic data sets that can be used for the early detection and individualized treatment of cancer. For example, he has developed a bioinformatics pipeline for the analysis of microarray data obtained from formalin-fixed, paraffin-embedded tissue samples for the identification of novel prognostic biomarker signatures.
Resource members contributed to the Cancer Biomedical Informatics Grid (caBIGÂ®) initiative of the National Cancer Institute as participants in the Vocabulary and Common Data Elements workspace of the caBIGÂ®. As part of the deployment phase, an operational caBIGÂ® node was installed on the 3CR network and the caTissue application was implemented as the inventory system for several projects collecting specimens.
As cancer research becomes increasingly data-driven, a growing data analysis bottleneck threatens to slow the rate at which new scientific discoveries are brought to the clinic. The Resource is dedicated to eliminating this using a wide-spectrum of techniques from the computational sciences to accelerate the pace of translational cancer research.
Gordon S. Okimoto, PhD
Lynne R. Wilkens, DrPH
Maj E. Earle