Association mapping for human diseases Karen Mohlke Ph D Department of Genetics University of North Carolina at Chapel Hill November 23 2009 Most human diseases are complex traits Familial but no obvious inheritance pattern Factors include genes environment behavior Complex interactions Common rare genetic variants Predisposition not causation Contributions to a complex trait risk to individual 2 100 risk to individual major gene a 90 80 100 90 minor gene a 80 70 70 60 60 50 40 30 20 10 minor gene b minor gene c exposure a 0 50 40 30 20 10 0 exposure b Evidence of a genetic component Higher concordance rate in monozygotic twins than dizygotic twins Higher relative risk to siblings parents and offspring than general population Verified genetic variants that influence susceptibility in model organisms or humans Gene mapping in populations Altshuler and Clark 2005 Science 307 1052 Genetic association study Cases risk allele Controls non risk allele Genetic association study Case control association study test for allele or genotype frequency differences in cases and controls Cases Controls n10 n01 n00 Logistic regression Chi squared test Trend test Logistic regression Cases Controls Chi squared test n11 n12 n11 n10 n02 n01 n00 Example chi squared test of association Observed Cases Controls Expected 65 121 186 5 45 141 186 5 110 3 262 7 372 row frequency column frequency sample size 5 3 372 55 Cases Controls 55 131 186 55 131 186 110 262 372 Example chi squared test of association Observed Expected Cases Controls 65 45 121 141 55 55 65 55 2 121 131 2 45 55 2 141 131 2 5 163 55 131 55 131 From a chi squared distribution with 1 degree of freedom P 0231 131 131 Single nucleotide polymorphisms SNPs GAAATAATTAATGTTTTCCTTCCTTCTCCTATTTTGTCCTTTACTTCAATTTATTTATTTATTATTAATATTATTATTTTTTGAGACGGAGTTTCACTCTTGT TGCCAACCTGGAGTGCAGTGGCGTGATCTCAGCTCACTGCACACTCCGCTTTCCGGTTTCAAGCGATTCTCCTGCCTCAGCCTCCTGAGTAGCTGGGACTACA GTCACACACCACCACGCCCGGCTAATTTTTGTATTTTTAGTAGAGTTGGGGTTTCACCATGTTGGCCAGACTGGTCTCGAACTCCTGACCTTGTGATCCGCCA GCCTCTGCCTCCCAAAGAGCTGGGATTACAGGCGTGAGCCACCGCGCTCGGCCCTTTGCATCAATTTCTACAGCTTGTTTTCTTTGCCTGGACTTTACAAGTC TTACCTTGTTCTGCCTTCAGATATTTGTGTGGTCTCATTCTGGTGTGCCAGTAGCTAAAAATCCATGATTTGCTCTCATCCCACTCCTGTTGTTCATCTCCTC TTATCTGGGGTCACCTATCTCTTCGTGATTGCATTCTGATCCCCAGTACTTAGCATGTGCGTAACAACTCTGCCTCTGCTTTCCCAGGCTGTTGATGGGGTGC TGTTCATGCCTCAGAAAAATGCATTGTAAGTTAAATTATTAAAGATTTTAAATATAGGAAAAAAGTAAGCAAACATAAGGAACAAAAAGGAAAGAACATGTAT TCTAATCCATTATTTATTATACAATTAAGAAATTTGGAAACTTTAGATTACACTGCTTTTAGAGATGGAGATGTAGTAAGTCTTTTACTCTTTACAAAATACA TGTGTTAGCAATTTTGGGAAGAATAGTAACTCACCCGAACAGTGTAATGTGAATATGTCACTTACTAGAGGAAAGAAGGCACTTGAAAAACATCTCTAAACCG TATAAAAACAATTACATCATAATGATGAAAACCCAAGGAATTTTTTTAGAAAACATTACCAGGGCTAATAACAAAGTAGAGCCACATGTCATTTATCTTCCCT ATTCATTTTCATAGTGGAAGAAATAAAATAAAGGTTGTGATGATTGTTGATTATTTTTTCTAGAGGGGTTGTCAGGGAAAGAAATTGCTTTTTTTCATTCTCT CTTTCCACTAAGAAAGTTCAACTATTAATTTAGGCACATACAATAATTACTCCATTCTAAAATGCCAAAAAGGTAATTTAAGAGACTTAAAACTGAAAAGTTT AAGATAGTCACACTGAACTATATTAAAAAATCCACAGGGTGGTTGGAACTAGGCCTTATATTAAAGAGGCTAAAAATTGCAATAAGACCACAGGCTTTAAATA TGGCTTTAAACTGTGAAAGGTGAAACTAGAATGAATAAAATCCTATAAATTTAAATCAAAAGAAAGAAACAAACTAAAATTAAAGTTATTATACAAGAATATG GTGGCCTGGATCTAGTGAACATATAGTAAAGATAAAACAGAATATTTCTGAAAAATCCTGGAAAATCTTTTGGGCTAACCTGAAAACAGTATATTTGAAACTA TTTTTAAAATGCAGTGATACTAGAAATATTTTAGAATCATATGTA A G 1 SNP 300 nt with frequency 1 10 million common SNPs in genome Amenable to high throughput genotyping Linkage disequilibrium C A A C A G C C A C C G T A A T A G T C A T C G Of 8 possible haplotypes if only 2 are observed then these SNPs are in linkage disequilibrium Markers tested in association analysis Goal is to test all putative causal alleles One SNP can act as a proxy for other SNPs if it is in strong linkage disequilibrium with the other SNPs Direct Indirect HapMap 2005 Nature 437 1299 Genome wide association GWA Benefits of GWA vs classical mapping More powerful vs linkage for common low penetrance variants Better resolution than linkage No need to select candidate genes Requirements of GWA Catalog of human genetic variants Low cost accurate method for genotyping Large number of informative samples Illumina Infinium genotyping assay Gunderson et al 2005 NatGen 37 549 GWA results for cholesterol triglycerides log10 p value log10 p value log10 p value 8 816 samples from 3 studies 2 3 million SNPs Willer 2008 Nat Gen 40 161 Significant lipid associations in stage 1 2 HDL C CETP 2e 57 LIPC 3e 20 LPL 4e 19 GALNT2 3e 14 LIPG 6e 12 ABCA1 1e 10 MVK MMAB 3e 8 LDL C APOE 3e 43 SORT1 6e 33 LDLR 4e 26 APOB 6e 22 PCSK9 4e 11 NCAN 3e 9 TG GCKR APOA5 LPL TRIB1 MLXIPL ANGPTL3 NCAN LIPC 6e 32 1e 26 4e 22 7e 13 2e 12 1e 10 3e 9 2e 8 Blue indicates previously known loci white indicates novel loci Willer 2008 Nat Gen 40 161 Example 1 Obvious candidate gene p1 2 6e 32 SORT1 sortilin 1 mediates endocytosis and degradation of lipoprotein lipase Nielsen JBC 274 8832 Example 2 Not obvious gene Pro92Ser p1 2 3e 9 NCAN nervous system specific proteoglycan involved in neuronal pattern formation Most likely affected gene and functional SNP ACACB Websites Genome Browsers genome ucsc edu www ensembl org SNP Database www ncbi nlm nih gov projects SNP