HMB321H1 Lecture Notes - Lecture 14: Genome-Wide Association Study, Snp Genotyping, Quality Control

45 views3 pages
user avatar
Published on 8 Oct 2011
School
UTSG
Department
Human Biology
Course
HMB321H1
9 MAR 2011
HMB321 L14: Genome Wide Association Studies (GWAS)
Common Disease / Common Variant Hypothesis
-GWAS are based on this CD/CVD hypothesis
-Where a common phenotype is associated with a common variant
-Thus, doing a GWAS in a large cohort of individuals may be able to find that “common” marker
-NIH GWAS (definition) = study of common genetic variation across entire human genome to identify
genetic associations with observable traits
GWA vs. Linkage
-Population study vs. Family study
-Whole genome vs. Pedigrees indicating candidate gene/regions
-Many SNPs vs. Few SNPs
-Can compare affected to unaffected (case-control) vs. Observed/expected P(x) of shared markers
-Complex vs. Single causative gene
-Not hypothesis driven (i.e. Search and then make hypothesis) vs. Hypothesis driven
What do GWAS study?
-Beyond finding disease genes => can be used to…
-Find gene variants with complex traits
-See slide for remainder
4 Components of GWAS
-(A) Select study population
-(B) Isolate DNA and genotype (SNP genotyping must pass quality control threshold)
-Make sure you accurately genotype
-Make sure each variant is a true variant (not something common to all)
-(C) Statistical significance
-Are geno/pheno truly associated?
-(D) Replicate in additional cohort + perform functional experiments
Sample selection + study design
- (1) Case Control
-Start with clinical records, i.e. Databanks => proper selection of groups is key (can alter results,
e.g. Women vs. Men in a study, e.g.2. Control group for CVD may include some non-clinical
symptom exhibiting people with CVD -- thus would need to further assess your control group
via family history of CVD to make sure that control has NO confounding possibilities)
- Find a way to separate 2 groups of patients = (a) disease (case) vs. (b) non-disease (control)
- Then genotype => compare
- (2) Cohort
- Look at clinical records => get cases full of diseased/prediseased (based on phenotype)
- Long-term study (because will have to look at onset of disease if pre-diseased)
- (3) Trio
- Look at the individual + the parents (hence: trio, 1 + 2P)
- Your null => no pheno/geno association (allele transmission = 50%)
-Key here is to accurately genotype to ensure you find true SNPs and not variants for all
-GWA design summary (look at the article given)
Unlock document

This preview shows page 1 of the document.
Unlock all 3 pages and 3 million more documents.

Already have an account? Log in

Document Summary

Where a common phenotype is associated with a common variant. Thus, doing a gwas in a large cohort of individuals may be able to nd that common marker. Nih gwas (de nition) = study of common genetic variation across entire human genome to identify genetic associations with observable traits. Can compare affected to unaffected (case-control) vs. observed/expected p(x) of shared markers. Not hypothesis driven (i. e. search and then make hypothesis) vs. hypothesis driven. Beyond nding disease genes => can be used to . (b) isolate dna and genotype (snp genotyping must pass quality control threshold) Make sure each variant is a true variant (not something common to all) (d) replicate in additional cohort + perform functional experiments. Start with clinical records, i. e. databanks => proper selection of groups is key (can alter results, Start with clinical records, i. e. databanks => proper selection of groups is key (can alter results, e. g. women vs. men in a study, e. g. 2.