A Survey of Four Disease Genetics Investigations using the Personalized Medicine Research Project

Tuesday, November 14, 2017 -
4:00pm to 5:00pm
Biotechnology Center Auditorium, 425 Henry Mall

Speaker Name: 

Steven Schrodi, PhD

Speaker Institution: 

Center of Human Genetics, Marshfield Clinic Research Institute




The Personalized Medicine Research Project biobank has many features making it an ideal source to conduct human genetics studies including a homogeneous population, rich medical record information, and high density genotype data. The aim of this seminar is to provide an overview of results from four disease genetics studies that my lab has recently conducted.  The molecular pathogenesis of virtually all common diseases remains enigmatic. Much of the work in my lab is predicated on the claim that a substantial repository of etiologic variants remains in the genome following standard screens interrogating additive effects from high-frequency alleles.

Mapping shared chromosomal regions in rheumatoid arthritis patients: Allelic and locus heterogeneity in common disease phenotypes are likely to be pervasive. We speculated that identifying shared chromosomal regions in very large extended kinships would enable the discovery of new disease variants. Using exome beadchip data, we conducted a scan of shared chromosomal regions in seropositive rheumatoid arthritis compared to sharing in sets of controls, revealing an interesting, statistically-significant signal on chromosome 3 harboring a candidate gene.
Gene-based compound heterozygote screening: Operating under the hypothesis that compromised-function diplotypes may drive pathology, the effects of which likely are concealed to traditional GWAS tests, we conducted a pilot study using exome genotype data (both phased and unphased), and subjected the results to a gene-based test. A pilot study of hemochromatosis confirmed the well-established HFE gene, and identified a novel gene encoding for a fibroblast growth factor. Extensive power calculations are shown supporting this approach and details of the statistical test shown.    

Bayesian Network-based prediction of early type 2 diabetes: To construct a molecular-based predictive model of early T2D, we conducted a study of PMRP individuals using samples obtained prior to a T2D diagnosis. A panel of metabolites, circulating cytokines and metabolic proteins, and GWAS-significant genetic markers were used in a Bayesian Network to produce a predictive model that outperformed glucose. 

A GWAS for baseline IL-17A protein levels: The IL-23/IL-17 axis is a pivotal network for autoinflammatory diseases such as ankylosing spondylitis, psoriasis and Crohn’s disease. Indeed, monoclonal antibodies targeting these cytokines have shown strong efficacy in the treatment of these diseases. We conducted a study of the genetic variants that underlie expression of the critically important inflammatory cytokine, IL-17A. The leukocyote-associated Ig-like receptor gene, LAIR1, was strongly associated with IL-17A levels. In addition, we show a highly significant increase in IL-17A with age.