The Statistical Genomics laboratory of Mayo Clinic researcher Brooke L. Fridley, Ph.D., focuses on novel statistical and analytical approaches to determine the genomic basis of complex diseases and traits. Dr. Fridley's Statistical Genomics lab is actively involved in the analysis of studies involving the genetic epidemiology of ovarian cancer and pharmacogenomics. These studies deal with both candidate genes and genomewide approaches, in addition to multiple types of genomic data, such as genotypic, methylation and mRNA expression. The Statistical Genomics lab is investigating methods for data integration, including bayesian methods and gene set analysis approaches. The lab recently started investigating and developing analysis methods for next-generation sequence data that will aid researchers in the interpretation of this high-dimensional data.
Here are two methods recently developed by the Statistical Genomics lab:
- Difference in minor allele frequency (DMAF) test. The lab recently developed this novel method for analysis of rare variants. The DMAF test allows combined analysis of common and rare variants and makes no assumptions about the direction of effects. DMAF can be used to analyze whole genes, or it can be applied in a sliding-window approach to localize the association signal in a region, using a step-down permutation approach to control type I errors with the testing of multiple windows. In an extensive set of simulations, Dr. Fridley's lab demonstrated that DMAF and other methods that pool data across individuals were found to outperform methods that pool data across variants when used to analyze traits with both risk and protective variants. The lab also found that the sliding-window DMAF method improved power compared with the whole-region analysis and that it was effective in fine mapping the causal subregion.
- PC-GM approach. The Statistical Genomics lab has completed research into a variety of gene-set analysis methods for both mRNA and genotypic data. In particular, the lab found that the combination of a principal component approach to conduct gene-level tests for association, followed by combining these p values using the Gamma method ENREF_3, a generalization of Fisher's method, to be powerful for gene set analysis of genotypic data from genomewide association studies (GWAS). Our lab refers to this new approach as the PC-GM approach.
- Statistical genetics and genomics
- Genetic epidemiology of ovarian cancer