Special Topics in Genomics ChIP-chip and Tiling Arrays

1
Special Topics in GenomicsChIP-chip and Tiling
Arrays
2
Traditional Method for Understanding
Transcription Regulation

• Very challenging for mammalian genomes

3
ChIP-chip Technology

• Chromatin ImmunoPrecipitation microarray
• Detect genome-wide in vivo location of TF and
  other DNA-binding proteins
• Can learn the regulatory mechanism of a
  transcription factor or DNA-binding protein much
  better and faster

4
Chromatin ImmunoPrecipitation (ChIP)
By Richard Bourgon at UC Berkley
5
TF/DNA Crosslinking in vivo
By Richard Bourgon at UC Berkley
6
Sonication (500bp)
By Richard Bourgon at UC Berkley
7
TF-specific Antibody
By Richard Bourgon at UC Berkley
8
Immunoprecipitation
By Richard Bourgon at UC Berkley
9
Reverse Crosslink and DNA Purification
By Richard Bourgon at UC Berkley
10
Amplification
By Richard Bourgon at UC Berkley
11
Genome Tiling Arrays
Arrays human genome Probes / Array Total Probes Probe Length Probe Resolution Price
Affymetrix 7 6M 42.0M 25mer 35 bp 2,000
Nimblegen 38 390K 14.8M 50mer 110 bp 30,000
Agilent 21 244K 5.1M 60mer 300 bp in genes 500 bp in intergenic 11,000
By Xiaole Shirley Liu at Harvard
12
Genome Tiling Arrays

• Affymetrix genome tiling microarrays
• Tile the genome non-repeat regions
• Chr21/22 tiling (earlier version) 1 million
  probe pairs (PM MM) at 35 bp resolution on 3
  arrays
• Whole genome 42 million PM probes on 7 arrays

PM CGACATTGATTCAAGACTACATACA MM
CGACATTGATTCTAGACTACATACA
Probes Chromosome
By Xiaole Shirley Liu at Harvard
13
Chromatin ImmunoPrecipitation (ChIP)
By Richard Bourgon at UC Berkley
14
ChIP-chip Array Hybridization

• Map high intensity probes back to the genome
• Locate TF binding location

ChIP-DNA
Noise
Probes Chromosome
By Xiaole Shirley Liu at Harvard
15
Identify ChIP-enriched Region

• Controls sonicated genomic Input DNA
• Often 3 ChIP, 3 Ctrl replicates are needed

ChIP
Ctrl
By Xiaole Shirley Liu at Harvard
16
Mann-Whitney U-testfor ChIP-region Detection

• Affy TAS, Cawley et al (Cell 2004)
• Each probe rank probes (either PM-MM or PM)
  within -500bp, 500bp window
• Check whether sum of ChIP ranks is much smaller

By Xiaole Shirley Liu at Harvard
17
TileMap (Ji and Wong, Bioinformatics 2005)
STEP 1 Compute a test statistic for each probe
to summarize probe level information
STEP 2 Combine probe level test statistics of
neighboring probes to help infer binding regions
18
Probe level test statistic empirical Bayes
approach
19
Combining neighboring probes
TileMap (MA) 1. Compute the probe level test
statistic t for each probe 2. Compute a moving
average statistic to measure enrichment 3.
Estimate FDR. TileMap (HMM) 1. Compute the probe
level test statistic t for each probe 2.
Estimate the distribution of t under H0 and
H1 3. Model t by a Hidden Markov Model, and
decode the HMM.
20
Shrinking variance increases statistical power
Moving Average
t-statistic, variance shrinking
t-statistic, canonical
Mean(X1)-Mean(X2)
21
Peak 2 (180bp) transgenics
Neural tube expression
Transgenics
22
Comparisons between TileMap and previous methods
cMyc ChIP-chip Data 6 IP 6 CT1 6 CT2 Gold
Standard Using GTRANS and Keles method to
analyze all 18 arrays
Test data 4 arrays, 2 IP vs 2 CT1 (s2r2)
TileMap-HMM (Ji Wong, 2005)
GTRANS or TAS (Kampa et al., 2004) 1. Set a
window 2. Perform a Wilcoxon signed rank test
for each window.
Keles et al. (2004) 1. Compute a t-statistic t
for each probe (no shrinking, two sample
only) 2. Rank probes by a moving average.
23
Shrinking variance saves money
Using non-shrinking method (Keles method) to
analyze all probes
Using shrinking method to analyze half of the
probes, i.e., reduce information by half
24
MAT(Johnson W.E. et al. PNAS, 2006)

• Model-based Analysis of Tiling arrays for
  ChIP-chip
• Goal
• Find ChIP-regions without replicates
• Find ChIP-region without controls
• Find ChIP-regions without MM probes
• Can analyze data array by array

By Xiaole Shirley Liu at Harvard
25
MAT

• Estimate probe behavior by checking other probes
  with similar sequence on the same array
• Probe sequence plays a
• big role in signal value
• Most of the probes in
• ChIP-chip measures
• non-specific
• hybridization

By Xiaole Shirley Liu at Harvard
26
Probe Behavior Model
Baseline on number of Ts
A,C,G,T Count Square
A,C,G at each position of the 25mer
25mer Copy Number along the Genome
By Xiaole Shirley Liu at Harvard
27
Probe Standardization

• Fit the probe model array by array
• Divide array probes to bins (3k probes/bin)
• Background-subtraction and standardization
  (normalization) on a single array

Model predicted probe intensity
Observed probe intensity
Observed probe variance within each bin
By Xiaole Shirley Liu at Harvard
28
Eliminate Normalization

• Probe log(PM) values before and after
  standardization
• If normalize before model fitting
• Predicted same ChIP-regions, although less
  confident

By Xiaole Shirley Liu at Harvard
29
ChIP-region Detection

• Window-based MATscore
• ChIP without Ctrl
• TM trimmed mean
• Multiple ChIP with multiple Ctrl
• More probes, higher t values in ChIP, less
  variance (fluctuation) ? more confident

By Xiaole Shirley Liu at Harvard
30
By Xiaole Shirley Liu at Harvard
31
Statistical Significance of Hits

• P-value and FDR cutoff
• P-value from MATscore distribution
• Estimate negative peaks under the same P value
  cutoff
• Regional FDR negative_peaks / positive_peaks

By Xiaole Shirley Liu at Harvard
32
MAT summary

• Open source python http//chip.dfci.harvard.edu/w
  li/MAT/
• Runs faster than array scanner
• Can work with single ChIP, multiple ChIP, and
  multiple ChIP with controls with increasing
  accuracy
• Use single ChIP on promoter arrays to test
  antibody and protocol before going whole genome
• Can identify individual failed samples

By Xiaole Shirley Liu at Harvard
33
Benchmark for ChIP-chip Target Detection(Johnson
D.S. et al. Genome Research, 2008)

• ENCODE Spike-in experiment both amplified and
  un-amplified
• Blind test Samples hybridized to different
  tiling arrays, predictions made before the key
  was released

ChIP 96 ENCODE clones, 2,4,8,...,256X enrichment
total chromatin DNA
Input total genomic DNA
34
Comparison of platforms
35
Comparison of algorithms
Combined Johnson D.S. et al. Genome Research 2008
with Ji H. et al. Nature Biotechnology 2008
36
MBR Microarray Blob Remover
By Xiaole Shirley Liu at Harvard
37
xMAN eXtreme MApping of oligoNucleotides

• http//chip.dfci.harvard.edu/wli/xMAN
• xMAN maps 42 M Affymetrix tiling probes to the
  newest human genome assembly in less than 6 CPU
  hours
• BLAST needs 20 CPU years BLAT needs 55 CPU days
• Probe TCCCAGCACTTTGGGAGGCTGAGGC maps to 50,660
  times in the genome
• Can map long oligos, and paired tag high
  throughput sequencing fragments
• Store the copy number information of every probe
• mXAN filters tiling array probes to ensure one
  unique probe measurement per 1 kb, improves peak
  detection

By Xiaole Shirley Liu at Harvard
38
CEAS Cis-regulatory Element Annotation System

• Data Analysis Button for Biologists

http//ceas.cbi.pku.edu.cn
By Xiaole Shirley Liu at Harvard
39
CisGenome(Ji H. et al. Nature Biotechnology,
2008)
Graphic User Interface
CisGenome Browser
Core Data Analysis Programs
40
Other applications of tiling arrays

• Transcriptome mapping
• MeDIP-chip
• DNase-chip
• Nucleosome localization
• Array CGH and copy number variation