|
Abstract:
|
Working with contingency tables, RNA-seq, single cell data often involves situations where the features available cannot be prespecified; thus the classical multivariate statistics theory does not apply. I will show how we can design nonparametric approaches that enable the integration and combination of different sets of measurements across samples. The approaches we have designed enable the integration of prior information about relations between features in the study of the human microbiome, the study of uncertainty in evaluating phylogenetic trees and the incorporation of spatial information in hierarchical models.
|
