Lecture 5: Multiple Comparisons
Last Lecture
Population substructure can cause
spurious association
Markers scatter throughout the genome
can be used to detect and correct for
substructure
Three commonly used methods
Genomic control
Structured associ
Lecture 8
Genotype Imputation &
Genetic Dissection of
Complex Diseases
1
http:/arjournals.annualreviews.org/doi/abs/10.1146/annurev.ph.34.030172.000305
Genotype Imputation
Haplotypes in Imputing Missing
Genotype Data
Haplotype structure can be used in
im
Challenges
Association analysis faces many
challenges, no matter what type of traits is
concerned
Spurious association caused by substructure
Multiple testing
Unknown genetic models
Small or moderate effect sizes of individual
variants/polymorphisms
Lecture 7: TDT and Extensions
1
Last Lecture
Methods to control for false discovery rate
(FDR)
Positive false discovery rate (pFDR)
q-value
2
This Lecture
Introduction to family-based association
test (FBAT)
Transmission disequilibrium test (TDT)
So
Lecture 3: Association Analysis in
Unrelated Individuals Single
Locus Analysis
1
Last Lecture
EM algorithm
ABO blood type
Infer haplotype phase
LD-based selection SNP Selection
ldSelect
Based on r2
2
Linkage Disequilibrium
Why LD is useful for mapp
Genotype and Rela-onship Error Iden-ca-on
Linkage programs will only run when the genotype data at the gene-c markers is
consistent with Mendelian segrega-on.
Two main sources of errors:
Pedigree Errors
Tru
Components of the Phenotypic Variance of
a Quantitative Trait
A quantitative trait is a continuous (and quantitative) variable that can be
measured in a population.
The total variance in a quantitative trait, termed the phenotypic variance,
can be partiti
Characterizing the Gene.cs of a Quan.ta.ve
Trait
How much of the trait varia.on is due to gene.cs, versus environment?
Heritability Method: Variance Component Analysis of Twins and/or Families
How many Quan.ta.v
Probability and Likelihood
If we know the values of a models parameters, we can compute the
probability of a set of data, given the parameter values.
A simple substitution polymorphism has two alleles, denoted A and a, with
allele frequency parameter valu
The Likelihood of a Pedigree
A genetic model is required that will describe the
probability of the unobserved disease mutation
genotypes in a family.
From the probability model, a likelihood can be
constructed and used to compute maximum
likelihood estima