Statement of Purpose

Programming Examples

Vegan Restaurants

Curriculum Vitae‎ > ‎

Masters Thesis

Transcription Factor Binding in a Familial Combined Hyperlipidemia

Weighted Gene Co-expression Network

by

Christopher L Plaisier

Master of Science in Bioinformatics

University of California, Los Angeles, 2009

Professor Steve Horvath, Chair

 

In our previous analyses of Mexican familial combined hyperlipidemia (FCHL) case/control fat biopsies we identified 28 co-expression modules using weighted gene co-expression network analysis. We identified that the URFA and midnightblue gene co-expression modules were associated with triglycerides. The URFA module was regulated by USF1 and the regulation of the midnightblue module was unknown. We hypothesized that we could identify the transcription factors regulating co-expression modules by screening for transcription factor binding sites enriched in the regulatory regions of a module’s genes. We identified five transcription factors regulating the URFA module (CEBPA, FOXD3, HLF, LHX3 and SOX5). The transcription factor CEBPA is known to be regulated by USF1, and was recently shown to harbor a variant affecting serum levels of triglycerides. We observed that the midnightblue module was regulated by the transcription factor GABPA. GABPA was also found to be enriched in the genes differentially expressed between FCHL cases and normolipidemic controls, thus linking this regulation back to disease. We then considered the amount of clinical trait variance explained by the gene co-expression modules. We hypothesized that utilizing pair-wise multiplicative interaction terms could account for additional trait variance. The interaction terms increased the fit of the model in most circumstances and were able to account for additional trait variance. Our final models were able to account for 33% of FCHL trait variance, and 36% of triglyceride trait variance. These findings extend previous analyses of Mexican FCHL case/control fat biopsies, and further demonstrate the utility of weighted gene co-expression networking method.

Comments