Biostatistics Shared Resource
BSR is a non-instrument based facility. BSR faculty interact with LCCC investigators to discuss scientific objectives in grant application and research projects, study design, biostatistics and data management and analysis plans. Masters level analysts are primarily involved in data analysis, study monitoring reports, and manuscript preparations.
Services offered by the shared resource include:
- Biostatistical support and consultation to LCCC investigators in study design and analysis. BBSR faculty and staff develop study designs and statistical analysis plans for clinical trials and observational investigations in collaboration with the study principal investigator (PI) and co-Is, perform corresponding sample size and power calculations, prepare randomization for clinical trials, assist the development of study case report forms (CRFs) to facilitate future analysis, develop and validate statistical models
- Biostatistics review and monitoring of clinical protocols through membership on the protocol review and monitoring committee (PRMC) and data safety monitoring committee (DSMC).
- Efficient design and analysis for translational in vitro and in vivo studies
- Consultation with investigators and data managers about utilizing their data for various types of statistical analyses and the preparation of reports and manuscripts. The BBSR also assists investigators in translating their computerized data into the format needed for analyses by various statistical software packages. Efforts in this area do not include the actual setting up of databases or entry of data for investigators.
- Additional specialized applications and collaboration utilizing advanced biostatistics and bioinformatics methods that have been developed by faculty with other funding mechanisms where necessary. BBSR faculty are at the forefront of developing new methods for big data analysis, transcriptome profiling, and innovative statistical designs. These new methods in natural language processing, computational biology modeling, novel clinical trial and animal studies design and analysis and causal inference are inspired by collaborations with members of the LCCC Programs and are utilized for the betterment of cancer research projects at LCCC.
- Biostatistics training of oncology fellows and medical students in trial design and analysis with structured seminars, course work and independent research projects, and free initial consultation for GUMC investigator in study design and analysis strategies
SAS and R are the systems utilized for most statistical analyses. The BSR also has access to statistical software packages that provide specialized capabilities. These computer programs facilitate study design and data analysis. The following is a partial list of the analytical software available on the BBSR’s machines: SAS, STATXACT, R, Matlab, Mathematica, STATISTICA, DSTPLAN (performs sample size and related calculations), NQuery (for sample-size computations), EGRET, EaSt, BART (program designed by D. Spiegelhalter for interim Bayesian analyses of clinical trials), BioConductor (R-based bioinformatics tools), BRB Array Tools (for the analysis of microarray data), dChip (for the analysis of Affymetrix microarray data analysis), CART (for classification and regression trees), IMSL (optimized C program library for numerical analysis), MOSEK (for functional optimization), Mplus, PDQuest (for analysis of two-dimensional gel images). All BBSR members also have access to Ingenuity Pathway Analysis for identifying networks from inferred differentially expressed genes. BBSR faculty and staff have developed open-source software including DeMix-Bayes (for tumor heterogeneity analysis), SysStat (design and analysis of drug combination studies), and EDRE (enhanced doubly robust causal inference).