APPLICATION FOCUS: ChIP-on-chip Gene Regulation

Chromatin immunoprecipitation (ChIP) is an effective tech- TABLE OF CONTENTS nique for enriching and identifying genomic DNA ChIP Applications 2 sequences bound by regulatory proteins, as well as iden- The ChIP-on-chip Toolbox 5 tifying sites of histone and DNA modification. The full Data Analysis Solutions 7 power of the application, however, is only realized when References 8 combined with very high-density, array-based detection. This allows interrogation of entire genomes at very high CRITICAL FEATURES resolution and has already revealed a wealth of essential Why are researchers choosing Affymetrix products for whole- information about regulatory activity in non-coding and genome ChIP-on-chip analysis? intergenic regions. ■ Lowest number of arrays possi- ble to cover the genome of It is no longer sufficient, or necessary, to limit regulatory studies to promoter re- interest gions or defined genomic loci. An unbiased, whole-genome approach reveals the full regulatory network activity of transcription factors and epigenetic modifications. ■ High-resolution probe design to Affymetrix' unique high-density tiling arrays accommodate 6.4 million features per pinpoint multiple data points array, allowing whole-genome (human or mouse) coverage on just a few arrays, de- per binding region, resulting in livering high-performance, cost-effective whole-genome ChIP-on-chip analysis. high accuracy and sensitivity ■ Easy-to-use tools for prelimi- nary and advanced data analysis

Enhancer

DNA methylation

Activator

Transcription Factor RNA Polymerase

DNA

Coding Region TATA Box

Figure 1: The total genome activity of transcription factors and epigenetic modifications are far more complex than previously predicted. APPLICATION FOCUS: ChIP-on-chip

Only Affymetrix high-density arrays Previous ChIP-on-chip experiments make whole-genome ChIP-on-chip ex- using focused arrays (as opposed to periments affordable and practical to whole- genome arrays) have traditionally Affymetrix whole-genome ar- conduct today—using just a single array analyzed candidate genes, annotated rays have already been used to for model organisms, and seven-array promoter regions or customized loci. sets for human and mouse. These These approaches have revealed only a map sites for transcription fac- whole-genome products allow you to partial picture of the total genome activ- tor binding, and DNA and his- analyze all coding and non-coding re- ity of key transcription factors and epi- tone modification, revealing gions for an unbiased view of the genetic modifications. genome, which is especially important key information about the func- With 25-mer probes spaced every 35 as non-coding regions contain several base pairs (bp) across an entire tions of various transcription enhancer/repressor binding sequences. genome, Affymetrix Tiling Arrays offer Single arrays from mouse and human regulators (including estrogen complete and unbiased genome cov- whole-genome sets are also available receptor [ER] and p53), and erage on a very small number of ar- for your initial assay development and rays. They offer the only practical, helping identify target genes optimization. Your initial experiments cost-effective and scalable approach give a realistic chromosome-wide view involved in disease progression to gene regulation analysis available of activity, and can be easily scaled up to date. and development—insights that for whole-genome analysis. Start your analysis with a single array could not have been made Recent reports have demonstrated that from a set, and confidently scale up to using focused arrays. epigenetic and transcription regulatory the entire genome, using the seven-array networks are far more complex than pre- set. No need to repeat large amounts of viously predicted, with individual tran- work—whole-genome analysis is afford- scription factors binding to thousands of Affymetrix Offers Unique Off- able and practical now. sites across the genome, and with much the-shelf Tools for Advanced of the regulatory activity located at distal Gene Regulation Analysis, sites far from promoter regions 5, 6. Which Provide Unrivalled Genome Coverage

■ Content: Over 6.4 million fea- tures per array for human and APPLICATIONS FOR IP-BASED ANALYSIS OF GENE REGULATION mouse sets The basics of Chromatin IP make it applicable to the study ■ Simplicity: Low number of arrays for human and mouse of any protein interaction with genomic DNA, or DNA or whole genomes, single whole- histone modification, provided an antibody exists that is genome arrays for other organ- isms specific for that protein or modification, and is suitable for ■ Resolution: 35 base pairs from immunoprecipitation. midpoint to midpoint of adja- cent oligos

Protein-DNA complexes are chemically comparison with non-enriched sam- ■ Specificity: 25-mer probes de- cross-linked, then purified from the cell ples for identification of all binding se- liver high sensitivity with a low and immunoprecipitated using an anti- quences—known and novel. incidence of false positives in body specific for the regulatory protein initial tests The ChIP-on-chip assay can be used to or modification under scrutiny. The examine gene regulation by both epige- DNA in the enriched fraction can be netic modifications, such as histone amplified using real-time PCR to con- and/or DNA modification, and by tran- firm the presence of expected se- scription machinery, such as transcrip- quences, and then applied to arrays for

2 Application Focus: ChIP-on-chip APPLICATION FOCUS: ChIP-on-chip

tion factor binding. ChIP-on-chip has sites. Dr. Kevin Struhl, an author of the also been used to study other protein- paper, discussed this work in the Octo- mediated processes such as DNA repli- ber 2005 issue of the Affymetrix Mi- “The combination of this unique cation and repair. croarray Bulletin. resource [whole-genome ChIP- Using a human whole-genome set Mapping Transcription Factor from Affymetrix, Yang, et al. (2006) ex- on-chip analysis at 35-bp resolu- Binding Sites amined the global binding behavior of tion] with gene expression data Recent evidence shows that many p63 and the evolutionary conservation transcription protein-binding sites in of p63 binding sites. sets serves to elucidate the the genome are actually in non-coding mechanisms underlying estro- regions once considered “junk” DNA, Mapping Histone Modifications and can sometimes be up to 100 kilo- gen-regulated gene expression Bernstein, et al. (2005) used Affymetrix bases (kb) from the gene they regulate. Tiling Arrays to map the location of two In addition, transcription factors are in breast cancer.” Carroll, et al. (2006) methylated lysines in the histones of more active than previously thought, embryonic stem cells. This modifica- with a high number of binding sites, tion pattern appears to be required to not always predictable from either silence developmental genes while canonical promoter regions or keeping the cells poised to differenti- “consensus motif” searches. ate. An interview with Dr. Bernstein Carroll, et al. (2006) published a discussing this work can be found in genome-wide map of estrogen recep- the October 2006 issue of the tor and RNA Polymerase II binding Affymetrix Microarray Bulletin. sites in breast cancer cells. The study Bernstein, et al. previously used tiling identified cis-binding sites in previously arrays to map histone methylation and unexplored regions of the genome, acetylation across non-repetitive por- and investigated cooperation between tions of human chromosome 21 and 22 transcription factors in estrogen signal- in both the human and mouse ing associated with breast cancer. Q- genomes. Bernstein’s team deter- PCR validation of 15 randomly selected mined that sites of trimethylation cor- binding sites confirmed ER recruitment Dr. Myles Brown in the October 2005 edition related with transcription start sites to all of the ChIP-on-chip identified of the Affymetrix Microarray Bulletin. and that methylation sites are strongly sites and none of the negative con- conserved between the two species. trols. Dr. Myles Brown, an author of Distribution of All TFBS Regions the paper, discussed this work in the Mapping DNA Methylation Pseudogene/ 5’ to a Known October 2005 issue of the Affymetrix Ambiguous Gene Microarray Bulletin (AMB), a magazine Zhang, et al. (2006) recently reported that highlights key microarray-based re- the first genome-wide analysis of DNA search from around the world. methylation in Arabidopsis thaliana at 35-bp resolution, using Affymetrix Tiling Cawley, et al. (2004) used Affymetrix Arrays. Using an anti-methylcytosine an- Tiling Arrays to map binding sites for tibody to precipitate methylated frag- the transcription factors Sp1, cMyc ments of the genome, the study found and p53. Thousands of previously un- that 19 percent of the Arabidopsis known transcription factor-binding sites Within 3’ or Flanking genome is methylated. They unexpect- Novel were identified—only about 22 per- to a Known Gene edly found that 35 percent of expressed cent of which were located at the 5’ Figure 2: Cawley, et al. (2004) Unbiased genes contain some methylation within end of protein-coding genes—disprov- mapping of transcription factor binding coding regions, while around 5 percent ing common assumptions about the sites along human chromosomes 21 and show methylation in promoter regions. location of transcription factor-binding 22 points to widespread regulation of noncoding RNAs. Cell 116, 499-509.

Application Focus: ChIP-on-chip 3 APPLICATION FOCUS: ChIP-on-chip

Hayashi, et al. (2006) employed a simi- methylated DNA-IP assay design and lar approach to map DNA methylation validation, array processing (using in the ENCODE regions of the human Affymetrix high-density, whole-genome “Our method recognizes DNA genome, identifying over 700 candidate Tiling Arrays) and advanced data analy- methylation with little bias for methylation sites. They also compared sis. An interview with Dr. Mary Warren, DNA methylation with histone H3 and Chief Scientific Officer of Genpathway, genomic location and therefore H4 acetylation, finding a reciprocal rela- discussing these services can be found is useful for comprehensive, tionship between the two. on the Affymetrix UserForum. These services include a wide range of ChIP high-resolution analysis of DNA Genpathway, an Affymetrix-approved assays covering gene expression and service provider, has developed a DNA gene regulation on a range of genomes. methylation.” Hayashi, et al. (2006) immunoprecipitation (IP) assay to en- For further information, visit www.gen- rich and detect methylated sequences pathway.com or contact your local on Affymetrix Human Whole-genome Affymetrix representative. Tiling Arrays. Genpathway recently ana- lyzed DNA methylation in MDA-MD- Mapping Other Protein-mediated Gene Categories 231 cells, an estrogen receptor- Biological Pathways negative breast cancer cell line Transcribed Region Lengronne, et al. (2004) used high- (http://genpathway.com/). On a single Promoter density tiling arrays to study how tran- array covering three chromosomes (8, scriptional activity determines where 70 11 and 12), they observed 997 methyla- sister chromosomes attach during cell 60 tion sites, of which only 223 were in or division, an important process in chro- 50 close to CpG islands (see Table 1), mosome segregation that goes awry 40 strengthening the argument for a in cancer. Cohesin, the protein that 30 whole-genome, unbiased approach to holds sister chromatids together, binds 20 epigenetic mapping using Affymetrix genomic DNA at unspecified sites, 10 whole-genome arrays. In addition, Methylated % Length and the unbiased and high-resolution 0 Genpathway’s IP approach using its op- ChIP-on-chip approach showed a strik- timized protocols results in the identifi- ing localization of cohesin to points of cation of methylated DNA regions with converging transcription. Dr. Frank higher sensitivity and specificity than Pseudogenes Uhlmann, principal investigator on the Non-expressed other methods. Known Expressed study, discussed the group’s findings Unknown Expressed Genpathway offers a uniquely compre- in the October 2005 issue of the Figure 3: Zhang X, et al. (2006) hensive package of chromatin-IP or Affymetrix Microarray Bulletin. Genome-wide High-resolution Map- ping and Functional Analysis of DNA Table 1: Locations of Peaks Representing Methylated DNA Methylation in Arabidopsis. Cell 126, 1-13. Peaks on Array F (Chromosomes 8, 11, 12) Number Percent

Total number on Array F 997

Extrapolated to entire genome 6070 “It will be important to examine In CpG Islands +/- 200 bp 233 23.4% the entire genome in ES cells

In Promoters (-1 to -10,000 bp from TSS*) 123 12.3% as well as to follow their In genes 584 58.6% fate during differentiation and

In exons +/- 100 bp 137 13.7% development.” Bernstein, et al. (2005) Unannotated (no gene within 10 kb) 216 22.0%

*TSS = Transcription Start Site1 Data provided by Genpathway

4 Application Focus: ChIP-on-chip APPLICATION FOCUS: ChIP-on-chip

THE AFFYMETRIX CHIP-ON-CHIP TOOLBOX The Affymetrix ChIP-on-chip toobox is designed to help you The Affymetrix ChIP-on-chip through every stage of the ChIP-on-chip workflow. The Toolbox ChIP-on-chip toolbox includes assay design and validation Designed to help you through every guidelines, reagents for sample labeling and preparation, stage of the ChIP-on-chip workflow. ■ Assay design and validation arrays for ChIP-on-chip, data analysis solutions, and com- guidelines munity support. This toolbox includes the Affymetrix ChIP- ■ Reagents for sample labeling and on-chip protocol, which serves as a “start-to-finish” preparation ■ Arrays for ChIP-on-chip guidance document on experimental design for a ChIP-on- ■ Data analysis solutions chip assay. ■ ChIP-on-chip community support

■ The protocol is available for ChIP-on-chip is a powerful technique for ■ Crosslinking and sonication condi- download from identifying sites of genomic regulation, tions www.affymetrix.com/support but can be a challenging task to under- ■ Availability of known positive control take for the first time. Affymetrix provides for IP enrichment QC the tools necessary to perform the assay, ■ and offers advice on assay optimization What negative controls will work and validation. best-Non-specific IP/Mock IP/Input? ChIP ASSAY GUIDELINES AND CHECKPOINTS ■ Amplification and fragmentation of The ChIP-on-chip Workflow Chromatin immunoprecipitation (ChIP) coupled with microarray PRE-IMMUNOPRECIPITATION technology provides a powerful new technique for enriching and identifying genomic DNA sequences bound by transcription factors and identifying sites of DNA and histone modification. Assay Design and Validation Guid- post-IP DNA Fix Cells to Cross-link DNA to Protein The number of cells should be 106-107; however, the This workflow describes a framework for performing the ChIP number may vary depending on cell line and application. assay, including cell fixation, sonication, immunoprecipitation, amplification, and hybridization, as well as several quality control (QC) checkpoints. lines Sonicate Samples to Shear Chromatin ■ Average length of DNA fragments should be approx- QC for enrichment pre- and post-am- imately 500 base pairs (bp). Sonication time and power may need to be adjusted. IMMUNOPRECIPITATION Immunoprecipitation Sonication QC Immunoprecipitate (IP) sonicated sample with selected plification Take small aliquot to reverse-crosslink and check ChIP-qualified antibody. Antibody amount may need to be The Affymetrix ChIP-on-chip protocol is fragment size on gel. Adjust sonication conditions to optimized. achieve optimal DNA fragmentation (approximately 500 bp). If successful, proceed to IP. Otherwise, adjust conditions to achieve proper fragment size. Immunoprecipitation Clean-up Couple precipitated DNA to Protein A beads and wash to available for download from www. enrich for IP’d DNA. Reverse-crosslink and treat DNA Ladder Sample with proteinase K. Purify IP’d material. Reagents for Sample Labeling and 1 2 Immunoprecipitation QC affymetrix.com. However, ChIP-on-chip Perform QPCR to check for enrichment. If a greater than 10-fold enrichment is not achieved, the IP should be Preparation repeated. Delta Rn vs. Cycle is a complex process, and each step of 1.0e+002 1.0e+001

1.0e+000 the protocol will need to be validated Affymetrix provides all of the reagents 1.0e-001 500 1.0e-002

1.0-e-003 by users in their own applications. Par- you need to label, purify and hybridize 1.0e-004 Cycle Number

Figure 2: QPCR data showing 10-fold enrichment. ticular areas for validation are: your sample post-amplification. Reagents AMPLIFICATION Figure 1: Sheared DNA from HL-60 cells following eight Amplify Enriched DNA pulses showed the optimal size range for immuno- Amplify enriched DNA fragments using random primer precipitation (majority of fragments between 300-500 bp). PCR (RP-PCR) or other DNA-based amplification for ChIP and amplification steps are technique. Spike dUTP’s into dNTP mix to enable fragmentation. Adjust the optimal number of cycles for RP-PCR if necessary to ■ avoid saturation. Hybridize 1-10 µg of amplified DNA to a single Verification of antibody utility for ChIP HYBRIDIZATION array; amount may vary depending on target complexity. user-supplied. ChIP-on-chip GeneChip® WT Double-Stranded DNA Terminal Labeling Amplification QC Kit (P/N 900812) can be used to fragment and label Perform QPCR to check if a greater than 10-fold (specificity, enrichment factor, etc.) amplified DNA using uracil-DNA-glycosylase (UDG) and enrichment is maintained. If lost, troubleshoot ApeI. Label amplified DNA target with terminal deoxynucleotidyl amplification. transferase (TdT). Hybridize sample to the arrays.

Affymetrix UserForum 3420 Central Expressway Santa Clara, CA 95051 USA www.affymetrix.com/userforum

Table 2: Reagents for ChIP-on-chip Applications The Affymetrix ChIP-on-chip workflow guides you through the key steps of the Reagent Product Code ChIP-on-chip assay. This workflow is available on the Affymetrix UserForum at: www.affymetrix.com/userForum/index.uf GeneChip® WT Double Stranded DNA Terminal 900812 Labelling Kit - 30 reactions

GeneChip® Sample CleanUp Module - 30 reactions 900371 GeneChip® Hybridization, Wash and Stain Kit - 30 reactions 900720

GeneChip® Control OligoB2, 3nM - 30 reactions 900301

Application Focus: ChIP-on-chip 5 APPLICATION FOCUS: ChIP-on-chip

Whole-genome Arrays for ChIP-on-chip Affymetrix Whole-genome Tiling Arrays for ChIP-on-chip analysis are available for human, mouse, Arabidopsis, C. elegans, S. cerevisiae, S. pombe and Drosophila.

Table 3: Whole-genome Tiling Arrays for ChIP-on-chip Array Product Code Pack Size

Human Tiling 2.0R Array Set (7 array set) 900772 1 Mouse 2.0R Array Set (7 array set) 900852 1

Arabidopsis Tiling 1.0R Array 900594 6 S. cerevisiae Tiling 1.0R Array 900645 6

S. pombe Tiling 1.0FR Array 900647 6 C. elegans Tiling 1.0R Array 900935 6

Drosophila Tiling 1.0R Array 900588 6

Table 4: Single Arrays from Human Whole-genome Set GeneChip® Array Product Code Pack Size

Human Tiling 2.0R Array A (chr 1, 6) 900779 6 Human Tiling 2.0R Array B (chr 2, 9, 19) 900780 6

Human Tiling 2.0R Array C (chr 23, 22, 21, X,Y,and mitochondria) 900781 6 Human Tiling 2.0R Array D (chr 4, 5, 18, 20) 900782 6

Human Tiling 2.0R Array E (chr 5,7, 16) 900783 6 Human Tiling 2.0R Array F (chr 8, 11, 12) 900784 6

Human Tiling 2.0R Array G (chr 10, 13, 14, 17) 900785 6

Table 5: Single Arrays from the Mouse Genome Set GeneChip® Array Product Code Pack Size

Mouse Tiling 2.0R Array A (chr 1, 9, 19) 900984 6 Human Tiling 2.0R Array B (chr 2, mitochondria, X, Y and unassigned) 900781 6

Mouse Tiling 2.0R Array C (chr 3, 7, 18) 900897 6 Mouse Tiling 2.0 R Array D (chr 4, 11, 17) 900895 6

Mouse Tiling 2.0 R Array E (chr 2, 12, 15) 900898 6 Mouse Tiling 2.0R Array F (chr 6, 8, 16) 900899 6

Mouse Tiling 2.0R Array G (chr 10, 13, 14) 900900 6

6 Application Focus: ChIP-on-chip APPLICATION FOCUS: ChIP-on-chip

Focused and Custom Arrays for ChIP-on-chip For research with a specific interest in promoter regions or ENCODE regions, for example, focused arrays are available. Custom Arrays are also often an attractive approach to ChIP-on-chip experiments. You can define any content, at any resolution, any strand detection, any application with a range of array formats accommodating from 38,000 to 6.5 million unique oligonucleotides per array.

Table 6: High-density Tiled Arrays for Promoter and ENCODE Regions GeneChip® Array Product Code Pack Size

Human Promoter 1.0R Array 900775 2 900776 6 900777 30

Human Chromosomes 21/22 2.0R Array 900936 6 ENCODE 2.0R Array 900937 6

Mouse Promoter 1.0R Array 900889 2 900890 6 900891 30

Data Analysis Solutions ■ Robust p-value calculation and false ■ Detect, display and report mapping discovery rate estimate sites of protein/DNA interaction Affymetrix offers its own tools for pre- liminary data analysis and visualization ■ Eliminates the need for normalization ■ Use both reference and user sam- ples as baseline of peaks, and can be freely downloaded ■ Freely available at: http://chip.dfci.har- from www.affymetrix.com/support. vard.edu/~wli/MAT/ ■ Create lists of genes and SNPs in binding regions ■ Tiling Analysis Software (TAS) will Partek® Genomics Suite perform basic raw data analysis to ■ Remove technical and biological produce signal and p-values for each Partek® Genomics Suite™ integrates ad- batch effects vanced statistics and interactive visuali- genomic probe on the array. ■ www.partek.com zation to reliably extract biological ■ Integrated Genome Browser (IGB) signals from noisy data. Designed for allows visualization of data taken Genomatix ChipInspector high-dimensional genomic studies, from TAS. You can select and load Partek GS is fast, memory efficient and A powerful program that extracts in- genomic annotations for cross-refer- will analyze large data sets on a per- formation from the expression level ence with ChIP-on-chip peak data. sonal computer. It supports a complete, of single probes using Affymetrix There are, however, a growing number easy-to-use workflow and is integrated tiling arrays. of other commercial and “freeware” with public genomic resources including ■ Increases the number of significant packages to help you with ChIP-on-chip GenBank®, NCBI GEO and NetAffx™. features while reducing the rate of data analysis. These include: false positives ■ Powerful statistics and interactive The MAT Algorithm from the Dana- visualization ■ Utilizes the world's largest database of alternative transcripts Farber Cancer Institute ■ Fast and memory-efficient ■ Assigns probes correctly to tran- A novel analysis algorithm to reliably de- ■ Easy-to-use workflows designed for tect regions enriched by ChIP on scientists scripts and genes Affymetrix tiling arrays developed by the ■ Eliminates interpolation/normaliza- ■ Identify regions of protein/DNA bind- laboratory of Dr. Xiaole Shirley Liu. ing in ChIP-on-chip studies tion problems ■ Available at www.genomatix.de ■ Can detect enriched regions from in- ■ Import and estimate mapping sites dividual ChIP samples of protein/DNA interaction

Application Focus: ChIP-on-chip 7 APPLICATION FOCUS: ChIP-on-chip

REFERENCES

6. Euskirchen G., et al. CREB Binds to Multiple 1. Bernstein B., et al. A Bivalent Chromatin Struc- AFFYMETRIX Inc. USA Loci on Human Chromosome 22. Molecular and ture Marks Key Developmental Genes in Embry- 3420 Central Expressway Cellular Biology 24(9):3804-14 (2004). onic Stem Cell Development. Cell 125:315-326 Santa Clara, CA 95051 USA (2006). 7. Hayashi H., et al. High-resolution DNA Methyla- Tel: 408-731-5000 (1-888-DNA- 2. Bernstein B., et al. Genomic Maps and Com- tion Mapping in Human Genome Using CHIP) parative Analysis of Histone Modification in Oligonucelotide Tiling Array. Human Genetics Fax: 408-731-5380 Human and Mouse. Cell 120:169-181 (2005). 120:701-11 (2007). www.affymetrix.com 3. Carroll J., et al. Genome-Wide analysis of Es- 8. Lengronne A., et al. Cohesin Relocation from trogen Receptor Binding Sites. Nature Genetics Sites of Chromosomal Loading to Places of Con- AFFYMETRIX, UK Ltd. 38(11):1289-97 (2006). vergent Transcription. Nature 430(6999):573-8 Voyager, Mercury Park, (2004). 4. Carroll J., et al. Chromosome-Wide Mapping of Wycombe Lane, Wooburn Green Estrogen Receptor Binding Reveals Long-Range 9. Yang A., et al. Relationships between p63 Bind- High Wycombe HP10 0HH Regulation Requiring the Forkhead Protein FoxA1. ing, DNA Sequence, Transcription Activity, and Bi- United Kingdom Cell 122:33-43 (2005). ological Function in Human Cells. Molecular Cell Tel: +44 (0) 1628 552550 24:593-602 (2006). Fax: +44 (0) 1628 552585 5. Cawley S., et al. Unbiased Mapping of Transcrip- [email protected] tion Factor Binding Sites Along Human Chromo- 10. Zhang X., et al. Genome-wide High-Resolution [email protected] somes 21 and 22 Points to Widespread Regulation Mapping and Functional Analysis of DNA Methy- of Noncoding RNAs. Cell 116 :499-509 (2004). lation in Arabidopsis. Cell 126:1-13 (2006). AFFYMETRIX JAPAN K.K. Affymetrix Japan K.K. Mita NN Bldg 16 Floor 4-1-23 Shiba Minato-ku, Tokyo 108-0014 Tel: *81-(0)3-5730-8100 Fax: *81-(0)3-5730-8101 [email protected] [email protected] www.affymetrix.com

Please visit our web site at www.affymetrix.com for interna- tional distributor contact information.

For research use only. Not for use in diagnostic procedures. Part No.702410-1

©2007 Affymetrix, Inc. All rights reserved.

Affymetrix® ™, , ™, GeneChip®, HuSNP®, GenFlex®, Flying Objective™, CustomExpress®, CustomSeq®, NetAffx™, Tools To Take You As Far As Your Vision®, The Way Ahead™, Powered by Affymetrix™, GeneChip-compatible™, and Command Console™ are trademarks of Affymetrix, Inc. All other trademarks are the prop- erty of their respective owners. Array products may be covered by one or more of the following patents and/or sold under license from Oxford Gene Tech- nology: U.S. Patent Nos. 5,445,934; 5,700,637; 5,744,305; 5,945,334; 6,054,270; 6,140,044; 6,261,776; 6,291,183; 6,346,413; 6,399,365; 6,420,169; 6,551,817; 6,610,482; 6,733,977; and EP 619 321; 373 203 and other U.S. or foreign patents.

8 Application Focus: ChIP-on-chip