Detection of Genes Influencing Chronic and Mendelian Disease Via Loss-of-Function Variation

<p>A typical human exome harbors dozens of loss-of-function (LOF) variants predicted to severely disrupt or abolish gene function. These variants are enriched at the extremely rare end of the allele frequency spectrum (< 0.1%), suggesting purifying selection against these sites. However, most previous population-based sequencing studies have not included analysis of genotype-phenotype relationships with LOF variants. Thus, the contribution of LOF variation to health and disease within the general population remains largely uncharacterized.</p> <p>Using whole exome sequence from 8,554 participants in the Atherosclerosis Risk in Communities (ARIC) study, we explored the impact of LOF variation on a broad spectrum of human phenotypes. First, we selected 20 common chronic disease risk factor phenotypes and performed gene-based association tests. Analysis of this sample verified two relationships in well-studied genes (<em>PCSK9</em> and <em>APOC3</em>) and identified eight new loci. Novel relationships included elevated fasting glucose in heterozygous carriers of LOF variation in <em>TXNDC5</em>, which encodes a biomarker for type 1 diabetes progression, and apparent recessive effects of <em>C1QTNF8</em> on serum magnesium levels. Next, we explored the effect of LOF variation on 308 small molecular metabolites, observing 8 significant genotype-phenotype associations. We highlight the relationship between serum histidine and<em> HAL, </em>a gene essential to histidine catabolism, demonstrating the biologically interpretability of associations with molecular metabolite targets. Finally, we explore the impact of LOF variation on a rare birth defect by comparing sequence from 342 unrelated left ventricular outflow tract obstruction (LVOTO) cases to ARIC sequence, identifying genes harboring case-exclusive LOF mutations. Comparison to an <em>a priori</em> list of cardiac candidate genes revealed 28 genes potentially related to LVOTO, including 22 not previously associated with a human disorder. Genotype validation in these samples revealed diverse inheritance patterns, including 9 confirmed de novo variants (<em>ACVR1</em>, <em>JARID2</em>, <em>KMT2D</em>, <em>NF1</em>, <em>NR2F2</em>, <em>PLRG1, SMURF1, TBX20</em>,<em> </em>and<em> ZEB2</em>).</p> <p>The analytical strategy presented here highlights the role of biologically-informed annotation on large-scale human genetic studies. The genes identified by these methods may have applications in disease prediction and drug development, and future genome studies will continue to refine our understanding of the scope of genetic variation affecting human health and disease.</p>

Name thesis:degree_name

Doctor of Philosophy (PhD)

Level thesis:degree_level

Dissertation (PhD)

Year dc:date.available

2015

Author dc:creator

• Li, Alexander H

Contributors dc:contributor

• Eric Boerwinkle
• Paul Scheet
• Hope Northrup

• genome
• exome
• loss-of-function
• rare variation
• gene
• disease
• Bioinformatics
• Epidemiology
• Genetics
• Genomics
• Medicine and Health Sciences

Repository record dc:identifier

https://digitalcommons.library.tmc.edu/utgsbs_dissertations/622

OAI identifier oai:identifier

oai:digitalcommons.library.tmc.edu:utgsbs_dissertations-1657