Statistical Methods for Biological and Relational Data

Methods for biological and relational data have pose challenges for statistical modeling. For biological data, gene expression data have high-dimensionality, and T-cell receptor (TCR) data under-sample receptor populations. For relational data, there are dependencies among the observations. This thesis outlines statistical methods for biological and relational data. The methods include classification, multiple testing and social networking. The models for classification are applied to gene expression data. The first method looks at variable selection to show the usefulness of sequential classification and regression trees to more advance methods. The second method uses Monte Carlo methods to calculate a rank for variable selection using supervised classification. Multiple testing methods are applied to gene expression and TCR data. The first method for gene expression looks at strong control of the familywise error rate without the assumption of the subset pivotality property, which is generally not met for gene expression data. For TCRs, the method extends the Poisson-lognormal model to the bivariate case to simultaneously analyze pairs of repertoires. The relational data uses social networking methods. The first uses exponential random graph models (ERGMs) with the application to political science. Solutions to two limitation of ERGMs, non-binary ties and longitudinal, are presented in examples. The last method proposes a latent position cluster model, an extension of latent class models that models clustering.

Name thesis:degree_name

Master of Science

Level thesis:degree_level

masters

Discipline thesis:degree_discipline

Biostatistics

Grantor dc:publisher

The Ohio State University

Year dc:date

2013

Author dc:creator

• Anderson, Sarah G.

Contributors dc:contributor

• Zhu, Hong

• Biostatistics
• gene expression
• T-cell receptors
• classification
• multiple testing
• relational data
• social networking

Statement dc:rights

• unrestricted
• This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws.

Language dc:language

English

Repository record dc:identifier

http://rave.ohiolink.edu/etdc/view?acc_num=osu1365441350

OAI identifier oai:identifier

oai:etd.ohiolink.edu:osu1365441350