Office hours Wednesday 3-4pm 304A Stanley Hall Review session 5pm Thursday, Dec. 11 GPB100.

Office hoursWednesday 3-4pm304A Stanley Hall

Review session5pm Thursday, Dec. 11

GPB100

Association vs. linkage

Strong, easy to detect, but rare in population;may not be reflective of common disease.Also, hard to collect family data.

Common but weak effects; need 1000’s of samples to detect.If no common cause, can fail.

Unrelatedindividuals

Relatedindividuals

Association vs. linkagesmall number of generations; individuals share big chunks of genome; can get co-inheritance between distant markers

many recombinations have happened since common ancestor;shared region is small; no co-inheritance between distant markers

So you need very high density of markers to get signal in an association study, but you get very high spatial resolution.

Association and admixture

Cases

Controls

=

=

At any one of these loci, Caucasian-like allele will be enriched in control samples.

Genotyping by array

Fig. 11.8

Expression microarrays

Fig. 1.13

Labeled DNA sample vs. labeled mRNA (cDNA) sample

Labeled DNA sample vs. labeled mRNA (cDNA) sample

(reverse transcription in the test tube)

Coding sequence array

Fig. 1.13


Fig. 1.13

say, with chemotherapeutic


Fig. 1.13


Fig. 1.13


Fig. 1.13

expression

expression

expressed

not expressed

Why measure expression of all genes at once?

“Regulon” expression response

http://www.mcb.mcgill.ca/~hallett/GEP/Lecture2/Image17.gif

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

(e.g. cortisol)

“Regulon” expression response



http://www2.kenyon.edu/Depts/BioEllipse/courses/biol114/Chap06/week08a_files/regulon.gif

Finding regulatory response on a large scale

Oligo expression array



transcripts

(soybean component)

Expression profiles



Time since drug administered

Expression profiles





Expression profiles





Each color is a regulon, or “cluster,” of co-regulated genes

Expression effects of cancer



Cancer classification



Cancer classification



Histological classification is finicky: can we do better?

Diagnosis via transcriptional profile






transcripts

patient samples






Diagnosis via transcriptional profileQuickTime™ and a

TIFF (Uncompressed) decompressorare needed to see this picture.

Natural variation among “normals”

Two human chromosomes differ at ~1/1000 bases.

96% of these differences are not in protein-coding sequence.Why?

Two human chromosomes differ at ~1/1000 bases.

96% of these differences are not in protein-coding sequence.Most protein coding mutations are deleterious;appear but are culled by natural selection.












Genetic variation in mRNA levels

ORFTFTF

G

kinaseTF

Genetic variation in mRNA levels

ORFTFTF

G

kinaseTF

Likely to be a complex trait.



Many mRNA differences at once



Linkage mapping of mRNA levels

“Black 6” mouse x “DBA” mouse



~10% mRNA levels significantly different



111 F2 progeny

…



111 F2 progeny

Microarrayeach F2 liver

…



111 F2 progeny


…

Genotypeeach F2



111 F2 progeny


…

Genotypeeach F2

Looking for linkage (coinheritance) between marker and mRNA level.

Marker is linked to polymorphism in expression regulation cascade

ORFTFTF

G

kinaseTF


ORFTFTF

G

kinaseTF

G

G


ORFTFTF

G

kinaseTF

G

G


ORFTFTF

G

kinaseTF

G

G One allele = high mRNA,the other = low mRNA


ORFTFTF

G

kinaseTF


ORFTFTF

G

kinaseTF


ORFTFTF

G

kinaseTF

mRNA level shows linkage to locus of polymorphic regulator(s).


ORFTFTF

G

kinaseTF

mRNA level shows linkage to locus of polymorphic regulator(s).

Locally acting polymorphisms









Polymorphism responsible for mRNA difference is at the locus of the gene itself


ORFTFTF

G

kinaseTF


ORFTFTF

G

kinaseTF









~25% of varying mRNAs are caused by locally acting polymorphism

Nonlocal polymorphisms



Nonlocal polymorphisms

One polymorphism in a key regulator can affect a regulon: 100’s of related mRNAs.

Clinical applications


111 F2 progeny


…

Genotypeeach F2

Measure fat pad each F2

QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.

QuickTime™ and aTIFF (Uncompressed) decompressorare needed to see this picture.







Colored curves = fat mass at different body locations


Finding polymorphism responsible for difference in macroscopic phenotype is hard


Finding polymorphism responsible for difference in macroscopic phenotype is hard

If mRNAs change too, can learn mechanism from known function of encoded proteins









Counts allele 1/allele 2, casesCounts allele 1/allele 2, controls

= 1.5



Marker predicts quantitative expression level = association

Can we map expression traits first, disease afterward?



Linkage of human transcripts

Linkage of human transcripts



Association of human transcripts



Association of human transcripts



linkage (families)

assoc (unrelated)

Association in multiple populations



Han Chinese and Japanese

European-Americans in Utah

Association in multiple populations



Office hours Wednesday 3-4pm 304A Stanley Hall Review session 5pm Thursday, Dec. 11 GPB100.

Documents

progeny slide

normals slide

cortisol slide

different slide

gpb100 slide

chemotherapeutic slide

mousexdba mouse slide

cancer classification