Presentation on theme: "Genetic research designs in the real world Vishwajit L Nimgaonkar MD, PhD University of Pittsburgh"— Presentation transcript:
Genetic research designs in the real world Vishwajit L Nimgaonkar MD, PhD University of Pittsburgh
Complex disorders: models of causation Genetic factors: Several genes induce cumulative, small but discrete effects + Environmental factors: etiological role / increased variability -No Etiological Factor Necessary or Sufficient -Formal proof dependent on statistical analyses
Factors influencing mapping efforts What is the phenotype? What polymorphisms are being used? What is the study design?
Key phenotype issues Is the phenotype heritable? –Proportion of risk due to genetic factors? –Proportion of risk due to an individual gene (# genes?) Familial aggregation does not necessarily prove genetic etiology Can the phenotype be evaluated reliably?
t Discrete (disease) Continuous (liability) 01 What is the phenotype? (L Almasy, PhD)
Phenotypes Qualitative (diagnostic status) –Clinically relevant –Difficulties in delineating ‘genetic’ phenotype Quantitative (‘endophenotype’) –Heritable –Differences between cases and controls –Differences between unaffected relatives & controls –Plausible role in pathogenesis, proximate to Dx
What polymorphisms? Single nucleotide polymorphisms: SNPs Repeat polymorphisms Insertions / deletions
What is the study design?
Gene mapping studies: concepts Examine correlation between genetic variation and trait of interest Significant correlation establishes genetic etiology
Human genome: 3 billion base pairs (estimated variations = 8,000,000 – 10,000,000) Problems 1. All genetic variations unknown 2. All variants can not be evaluated
Marker A2 Marker A1 Marker B1 Marker B2 *Mutation* Haplotype 1 Gene mapping concepts control case
Recombination based gene mapping Generations: Transmission Of Disease Gene Ill Individual n Transmission Of Normal Gene Healthy Individual
Linkage / Association Linkage Association generations founder
What is the study design? POSITIONAL CLONING Step 1: Identify large shared chromosomal segments among cases within families (LINKAGE) Step 2: Narrow the shared region using cases and controls (ASSOCIATION).
Linkage: haplotype sharing
Related issues Ascertainment and recruitment! Power: more is better! ‘much, much more’ preferred Design modification –Two stage design (accept lower lod cutoffs) –Covariate based analyses
ASP analysis Convenient design Concerns –Truncation of family size due to morbidity –‘True’ sibling recurrence risk –Uncertain paternity –Twinning Power: n = 400 ASPs; power > 80% for λs = 3.0 (LOD = 3)
Sample size required for 80% power to detect linkage to a QTL at a LOD of 3 (Almasy et al.)
Associations at the population-level Generations: Transmission Of Disease Gene Ill Individual n Transmission Of Normal Gene Healthy Individual
Factors influencing associations Sample selection & size Population history (fitness, drift, migration) Features of mutations (no, age, frequency) Features of markers (informativeness, LD) Number of comparisons Ethnic admixture
Family based associations (haplotype relative risk) A, CB, D A, BC, D Hypothetical control
Family based associations Recruitment expensive Ascertainment may be biased Easier than multiplex pedigrees Power: Issues –Uncertain paternity –Genotyping errors –Power diminishes for case-parent duos
‘Novel’ designs Cytogenetic abnormalities Pooled DNA analyses