Finding Disease Genes

Question 1

why is it hard to predict the disease base on gene hard?

Answer 1

Low Odds Ratio

Question 2

Odds Ratio

Answer 2

Risk metric

OR= risk of disease in given gene variant/ risk of disease without given gene variant

Question 3

two approaches to find gene?

Answer 3

function to gene (old way)
gene to function (new, more effective)

Question 4

how do we map the gene?

Answer 4

DNA markers

Question 5

3 types of DNA markers

Answer 5

• microsatellites
• single nucleotide polymorphisms (SNPs)
• copy number variations (CNVs)

Question 6

microsatellites

Answer 6

simple sequence repeats
multi allelic
used for foresnsics
1/30000 bp

Question 7

Single nucleotide polymorphisms (SNPs)

Answer 7

• occurrence/allele frequencies differ in different ethic groups/populations
• occurs in haplotype

Question 8

haplotype

Answer 8

• macro patterns of polymorphic genotypes on the same chromosome
• recombination sets it

Question 9

haplotype block

Answer 9

• linkage disequilibrium:
• marker alleles within blocks tend to be co inherited
• recombination within block is uncommon

Question 10

Copy Number Variants (CNVs)

Answer 10

common genomic deletions

*most not causal for human disease.

Question 11

Hypothesis driven approach, 1

- candidate gene DNA sequencing

Answer 11

(hit from GWAS or other mapping method) works for mendalion)

- most hypotheses wrong

Question 12

Hypothesis driven, 2
- Candidate gene association study
(test gene/causal variant indirectly)

Answer 12

• most common genetic study
• depends on biological(priori biological hypothesis) positional candidate (positional hypothesis)
• most powerful for common risk alleles with small to moderate ORs (complex, polygenic traits)
• most priori biological hypotheses—> wrong
• ***fetal flaws lead to false positives

Question 13

Candidate Gene Association Study– concepts

Answer 13

1. causal disease variation in candidate gene tagged by local hyplotype of polymorphic DNA markers
2. depends on linkage disequilibrium

Question 14

candidate gene association study–approach

Answer 14

case control study design
1. genotype marker in candidate gene (for both experimental groups and controls)

2. compare allele frequencies btw the two groups

Question 15

Facts about genetic assoication studies

Answer 15

1. simple
2. big size number of cases and controls (hundreds)
3. simple stats (p value, chi square, fisher exact test)
4. if multiple variants tested, must have multiple testing correction
5. no causation
6. imply linkage disequlibrium with mutation
7. ** ALMOST ALWAYS YIELD FALSE POSITIVE.

Question 16

case control genetic association study’s false positive’s reasons

Answer 16

1. multiple testing correction not including all tests done by others including never published one.
2. background genetic variation influence due to imperfect matching of cases and controls.
   • even in homogenous population stratification can give false positive.

Question 17

Hypothesis free approach, 1

Answer 17

Genetic linkage analysis

• search genome for segments co-inherited disproportionately with disease through multiplex families
• affected relatives within family share disease susceptibility genes “identical by descent”
• powerful for mendelian
• less powerful for complex traits

Question 18

centi Morgan (cM)

Answer 18

the unit of genetic distance/recombination

statistical measure is LOD (log of odds) score

significance level: LOD >_ 3.0 is considred proof of linkage/gene localization

Question 19

Hypothesis free approach, 2

Answer 19

Genome Wide Association Studies (GWAS)

• same as case control association study except so many SNPs across entire chromosome.
• look for SNPs with significantly different alleles frequencies in cases versus controls
• can do right multiple testing correction
• entire genome representation fixes the population stratification

Question 20

Genome Wide Association study facts

Answer 20

• very large number of SNPs tested gives huge multiple testing problems (requires a lot of cases and control)
• significant assocaition—>needs confirmation by independent replication (follow up association study of specific SNPs)
• no hypothesis needed
• can discover new genes
• most effective for common alleles with small to moderate ORs (effective size)

Question 21

Combined hypothesis based and hypothesis free approaches

Answer 21

deep re sequencing

• biological candidate genes
• GWAS signals
• full genome or exome sequencing
• causal variants from non pathological variation
• prioritize for follow up functional analysis

Question 22

Exom/genome sequencing in mendelian disorder

Answer 22

• lots of genomic noise

- going to be fairly routine clinical test for us?