Multiple-component Analysis in Multivariate Additive Regression Modeling Applied to NMR biomolecular Studies
Multiple-component analysis in multivariate additive regression models is a new research statistical problem that arose with the recent technological advancements in biomolecular studies and proteomics. This model formulation is novel and it applies to many applications where the data are a sum of identifiable components, where each component may correspond for example, to a different protein in proteomics experiments, or on another hand, benign and malignant cancer lesions in CT experiments.
In this presentation, I will introduce this general modeling framework applied to the analysis of biomolecular studies using Nuclear Magnetic Resonance (NMR).
Specific research topics covered in this talk are:
1. To shed light into the potential of wavelet-based noise reduction and multiple component identification in biomolecular NMR data analysis.
2 To explore an estimation algorithm designed to overcome the curse of dimensionality arising in the proposed multi-dimensional modeling framework;
3. To introduce inferential procedures including hypothesis tests for non-identifiability and mixture of components.
These statistical methods will be discussed in the context of the motivating application, NMR biomolecular data analysis.