eQTL mapping and inherited risk enrichment analysis : a systems biology approach for coronary artery disease

Despite extensive research during the last decades, coronary artery disease (CAD) remains the number one cause of death, responsible for near 50% of global mortality. A main reason for this is that CAD has a complex inheritance and etiology that unlike rare single gene disorders cannot fully be understood from studies of of genes one-by-one.In parallel, studies that simultaneously assess multiple, functionally associated genes are warranted. For this reason we undertook the Stockholm Atherosclerosis Gene Expression (STAGE) study that besides careful clinical characterization and genome-wide DNA genotyping also assessed the global gene expression profiles from seven CAD-relevant vascular and metabolic tissues. In paper I, we used STAGE to develop a bioinformatics tool for efficient eQTL mapping called kruX based on Kruskal-Wallis statistics test. kruX excels in detecting a higher proportion of nonlinear expression quantitative expression traits loci (eQTLs) compared to other established methods. This tool was developed for Python, MATLAB, and R and is available online. In paper II, we applied kruX to detect eQTLs across the seven tissues in STAGE and assessed their tissue specificity. A tool for analyzing inherited risk enrichment was also developed assessing CAD association (i.e., risk enrichment) of STAGE eQTLs according to genome-wide association studies (GWAS) of CAD. We found that eQTLs active across multiple vascular and metabolic tissues are more enriched in inherited risk for CAD than tissue-specific eQTLs. In paper III, we integrate the analysis of STAGE data with data from GWAS of CAD to identify 30 regulatory-gene networks causal for CAD. In paper IV, we again used kruX to investigate STAGE eQTLs for three established candidate genes in CAD and atherosclerosis (ALOX5, ALOX5AP, and LTA4H). In addition, we used the Athero-Express dataset of genotype and atherosclerotic carotid plaque characteristics to further elucidate the role of these genes in atherosclerosis development. In sum, in this thesis report we show that by integrating GWAS with genetics of gene expression studies like STAGE, we can advance our understanding from the perspective of multiple genes and gene variants acting in conjunction to cause CAD in the form of regulatory gene networks. This is done through developing new bioinformatics tools and applying them to disease-specific, genetics of global gene expression studies like STAGE. These tools are necessary to go beyond our current limited single-gene understanding of complex traits, like CAD. List of Publications This thesis is based on the publications listed below. I Jianlong Qi, Hassan Foroughi Asl, Johan Björkegren, Tom Michoel. kruX: matrix-based non-parametric eQTL discovery. BMC Bioinformatics 2014 Jan 14; 15:11.doi: 10.1186/1471-2105-15-11 II Hassan Foroughi Asl, Husain A Talukdar, Alida SD Kindt, Rajeev K Jain, Raili Ermel, Arno Ruusalepp, Khanh-Dung H Nguyen, Radu Dobrin, Dermot F Reilly, Heribert Schunkert, Nilesh J Samani, Ingrid Brænne, Jeanette Erdmann, Olle Melander, Jianlong Qi, Torbjörn Ivert, Josefin Skogsberg, Eric E Schadt, Tom Michoel, Johan L.M. Björkegren. Expression quantitative trait Loci acting across multiple tissues are enriched in inherited risk for coronary artery disease. Circulation Cardiovascular Genetics. 2015 Apr; 8(2):305-15. doi: 10.1161/CIRCGENETICS.114.000640 III Husain A. Talukdar, Hassan Foroughi Asl, Rajeev K. Jain, Raili Ermel, Arno Ruusalepp, Oscar Franzén, Brian A. Kidd, Ben Readhead, Chiara Giannarelli, Jason C. Kovacic, Torbjörn Ivert, Joel T. Dudley, Mete Civelek, Aldons J. Lusis, Eric E. Schadt, Josefin Skogsberg, Tom Michoel,and Johan L.M. Björkegren. Cross-Tissue Regulatory Gene Networks in Coronary Artery Disease Cell Systems. 2016 March; 2(3):196-208. doi: 10.1016/j.cels.2016.02.002. IV Sander W van der Laan*, Hassan Foroughi Asl*, Pleunie van den Borne, Jessica van Setten, ME Madeleine van der Perk, Sander M van de Weg, Arjan H Schoneveld, Dominique PV de Kleijn, Tom Michoel, Johan L.M. Björkegren, Hester M den Ruijter, Folkert W Asselbergs, Paul IW de Bakker, Gerard Pasterkamp Variants in ALOX5, ALOX5AP and LTA4H are not associated with atherosclerotic plaque phenotypes: the Athero Express Genomics Study. Atherosclerosis. 2015 Apr;239(2):528-38. doi: 10.1016/j. atherosclerosis. 2015.01.018.