Introduction to Affymetrix SNP technology

Jonathan Pevsner, Ph.D.SNP group at the Genome Cafe

July 27, 2005

Overview of chromosomes and DNA

Human genome: 23 pairs of chromosomes

22 autosomal pairs1 pair of sex chromosomes

Scales of DNA

Chromosome 10: 135 Mb (983 genes)

Chromosome 21: 46 Mb (337 genes)

Chromosome 1: 245 Mb (2580 genes)

Human genome: 3 billion base pairs

Scales of DNA (four bases: A,C,T,G)

1 base pair: may differ (single nucleotide polymorphism, “SNP”)between individuals

800 base pairs: length of typical DNA sequencing read800 base pairs: convenient size of PCR product1,000 base pairs: size of a typical protein-coding region8,500 base pairs: median SNP interval on 100K Affy chip100,000 base pairs (100 kilobases): size of a typical gene200,000 base pairs: size of a typical BAC clone3 million base pairs (3 Mb): smallest

size deletion easily visible with conventional cytogenetics

150 Mb: typical chromosome3,000 Mb (3 Gb): human genome

proteinRNA

DNA

Central dogma of molecular biology

Across the genome, there are four possible SNP calls:[1] homozygous (AA)[2] homozygous (BB)[3] heterozygous (AB)[4] no call

In a deleted region, the possiblecalls are A, B, or no call.The computer interprets thesepossible calls as AA, BB, orno call. AA or BB SNP calls are homozygous.

Across the genome, there are four possible SNP calls:[1] homozygous (AA)[2] homozygous (BB)[3] heterozygous (AB)[4] no call

genomic DNA (250 ng)

cell line (control or experimental)

digest (XbaI or HindIII)

ligate adapters

PCR amplification

fragment and end-label

hybridize

wash

scan

acquire raw intensity values

normalize

assess SNPintensity values

assess SNPLOH values

Samples

Normalfemale

Normalmale

Affectedindividual

DNA cells

GeneChip DNA Analysis Software (GDAS):.CEL file

GeneChip DNA Analysis Software (GDAS)

.CHP file

.CEL file

AB

AABB

confidencescore

dbSNP IDcall

callclassification

“no call”

callclassification

confidencescore

For normal males, there shouldbe three possible SNP calls for the X chromsome: [1] homozygous (A is interpretedby the algorithm as AA)[2] homozygous (B is interpretedby the algorithm as BB)[3] no callHeterozygous (AB) calls shouldonly happen as an error

For normal females, there are four possible SNP callsfor the X chromosome:[1] homozygous (AA)[2] homozygous (BB)[3] heterozygous (AB)[4] no call

Affymetrix Confidence Score

0.001 0.01 0.1 1

No Call

CalledHeterozygote

CalledHomozygote

0.25

Pro

babi

lity

of

a w

rong

cal

l

n = 7651

n = 52

n = 585

0.001

0.010

0.100

Probe Set: an Affy idChromosome: 1-22, X, Y, unassignedPhysical Position: varies by build!Call: AA, BB, AB, or no callSPA_CN: copy numberSPA_pVal: p value for copy number changeGSA_pVal: p value for copy number changeLOH: -log10 p value for loss of heterozygosity

Case #1 SNP data

Deleted region on chr7:Only expected calls: AA, BB, no call

Possible calls: AA, BB, AB, no call

Case #2: SNP arrays - chr 2

Case #2: SNP arrays - chr X

Web site

140

280

420

560

700

consecutive homozygosity size (number of calls)

cou

nt

Web site

chromosome caseho

moz

ygos

ity s

tret

ch

childH8

2298

2299

1054

2283

2284

1928

LT

MT

parent1

parent2

2287

2297

3387

2287

2297

3387

1 2 3 4 5 6 7

chromosome

hom

ozyg

osity

str

etch

A

B

con

secu

tive

hom

ozy

gote

s

physical location, chromosome 22 (megabases)

10 20 30 40 50

nor

ma

lize

d

dye

ra

tios

50

40

30

20

10

0

1.4

1.2

1.0

0.8

0.6

0.4

centromere

Integration of SNP chip and genomic microarray data