Needs Assessment and Project Design (Consultation)
Consultation and assistance is provided for experimental design, data mining, downstream statistical and biomedical informatics data analyses, as well as data interpretation of disparate data sets to determine the underlying genetic and environmental mechanisms that lead to cancer disease phenotypes. Helping investigators clearly define their biomedical informatics needs is a first step towards successful effective project execution. Integration of disparate data sets to support specific cancer projects requires an assessment of the types of data required, hypotheses to be tested, tools needed to interface the data resource and collaborators on the project. Understanding these key issues is essential to determining the scope of the work, time requirements, expertise and customized resources needed. Recommendations are made on storage methods for data, as well as database and data analysis tools. Security and privacy concerns are assessed to assure these data are maintained and protected, as required.
Development of Biomedical Computational Tools
The Biomedical Informatics Shared Resource Facility helps develop computational tools for investigator-initiated projects with the assistance of Research Information Technology (IT). Services provided by Research IT and the Biomedical Informatics Facility include selection of data models to accurately represent the relationships of data, database design and programming; designing hardware architectures for electronic collection of data; programming web-enabled user interfaces; development of new data visualization tools; and provision of the necessary programming skills to integrate open-source and commercial software. The Biomedical Informatics Facility also provides the expertise to integrate data resources and tools into cancer-related data grids, including caBIG™.
Integration of Data
This service area includes integrating experimental measurements with appropriate clinical metadata for cancer research projects that utilize web-based technologies. A key to developing useful computational tools to support cancer research is the creation of data models that properly represent molecular and clinical data (particularly incorporating social, environmental and longitudinal data) to support, for example, clinical trials, molecular diagnostics and patient management. Each investigator-initiated project requires flexible, scalable and portable data resources that are based on standardized data models, vocabularies and metadata schemas. The Biomedical Informatics Facility, working closely with Research IT, assists researchers by providing flexible data models that deliver the necessary architecture so that when new data types emerge in the quick-paced fields of genomics and proteomics, these data can be integrated into the data resource and linked to previous data types. Additionally, data resources are designed to provide scalability to accommodate the enormous amounts of data generated by high-throughput technologies and to incorporate portability by using standards such as XML for data exchange.
Evaluation of Software Tools
The facility assesses both open-resource and commercial software for investigators. Software for bioinformatics, databases, end-user interfaces, operating systems, data integration and data analysis is evolving at an extremely fast pace. By keeping abreast of new products, collecting product information and testing software, the facility is able to help cancer researchers select the appropriate tools, as well as implement new software, as needed.
Analysis of Biomedical Data
In collaboration with the Biostatistics Division led by Michael J. Schell, Ph.D., the facility provides data analysis when it is impractical or beyond the knowledge of the researcher. Examples of the types of data analysis include linkage disequilibrium, haplotype determination, hierarchical and non-hierarchical clustering, principal component analysis, Hardy-Weinberg calculations, logistical regression models and power calculations. Expert consultations assist cancer investigators by supporting sophisticated data analyses for proteomics, microarrays, genotyping and bioinformatics. When possible, the Biomedical Informatics Facility provides hands-on training for investigators on the methods and tools needed to analyze their own data.
Education and Training
Training and education programs such as workshops and seminars, as well as presentations at Cancer Center research group seminars or retreats, are an important component of the Biomedical Informatics Facility services to end-users. Educational programs provide information on the underlying data structures, methods for viewing and querying the data resources and downstream analyses to find concordance between phenotype and genotype data. The Biomedical Informatics Facility also works closely with other Cancer Center Shared Resources to ensure that its educational programs include information on biostatistics analysis methods, data mining techniques (also, KDD – knowledge discovery from databases) and molecular data analyses. In addition to these activities, the Facility Director and staff are available for one-on-one sessions with investigators on the use of data sources, how biomedical informatics services can support their research, as well as analysis tools available.
© Copyright 1996 - 2013 H. Lee Moffitt Cancer Center & Research Institute