DNA Methylation

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

1.1 DNA Methylation

DNA methylation is a form of post-replicative or epigenetic modification of DNA that influences cellular functions by altering the gene expression but do not involve changes in the underlying genomic DNA sequence. DNA methylation is critical for normal mammalian embryonic development, gene silencing, X chromosome inactivation and imprinting. It acts as a cue for strand specificity in DNA replication and repair and could also serve as a defense mechanism to silence transposable elements. This post-synthesis modification occurs only at carbon number 5 in CpG dinucleotides of which between 70 to 80% in humans. Abnormal DNA methylation patterns are closely associated with most cancers in mammals (Knowles and Selby 2005; Szyf 2005).

1.2 CpG Islands

In humans, the distribution of DNA methylation are not random throughout the genome, instead the methylated DNA is localized to discrete regions rich in repetitive DNA, transposable elements, imprinted domains, and some inactive X chromosomes in females. At these regions, DNA methylation may act to suppress the transcription, transposition and recombination of certain genes.

DNA methylation is primarily involved in the process of tumorigenesis, whereby the tumour cells show global losses of methylation from repetitive sequences and region-specific gains in methylation, especially within the CpG-rich gene regulatory regions (Esteller 2005).

Unmethylated CpGs are grouped in clusters termed "CpG islands". These CpG islands comprises of around 50% of all unmethylated CpGs in the genome. The methylation of the CpG islands may alter gene expressions by its influence on chromatin organization. In many disease processes such as cancer, gene promoter CpG islands acquire abnormal hypermethylation, which results in heritable transcriptional silencing (Isaacs and Rebbeck 2007).

CpG islands are usually found in close proximity to the promoter regions of genes that are involved in general cellular functions termed housekeeping genes or other genes that are frequently expressed in a cell. Methylation of the promoter regions can silence the associated gene expression efficiently. At these regions, the CG sequences in active genes are not methylated whereas the CG sequences in inactive genes are usually methylated to suppress their expression.

The methylated cytosine may be deaminated to thymine spontaneously. Unlike cytosine to uracil mutation that can be repaired efficiently, cytosine to thymine mutation can only be repaired by mismatch repair, which is highly inefficient.Thus the methylated CG sequence will be converted to a TG sequence. This explains the lack of the CG sequence in inactive genes.

Each and every cell has its own unique methylation pattern so that a unique set of proteins may be expressed to perform cellular functions specific to the cell. Hence during cell division, the methylation pattern should also pass on to the daughter cell. The enzyme DNA methyltransferase, which can methylate only the CG sequence paired with methylated CG, is responsible for this.

Diagram 2.1: Inheritance of the DNA methylation pattern. The DNA methyltransferase can methylate only the CG sequence paired with methylated CG. The CG sequence not paired with methylated CG will not be methylated. Hence, the original pattern can be maintained after DNA replication.

1.3 Types of Methylation

There are two types of methylation namely hypomethylation and hypermethylation. Common characteristics of mammalian tumor cells include global hypomethylation and regional hypomethylation and hypermethylation.

Regional hypomethylation is associated with cancer via the activation of proto-oncogene, which is usually methylated and expression silenced within the genome. DNA hypomethylation is also associated with tumor progression. Hypomethylation of specific DNA sequences particularly in repetitive DNA sequences may act as a marker for tumor progression as can hypermutation of unique DNA sequences. Cancer-associated DNA hypomethylation can occur without DNA hypermethylation. However a possible setback of hypomethylation is the association with genomic instability, resulting in a loss of heterozygosity (LOH). LOH is defined as the loss of one allele at a specific site on the chromosome (locus), which can thus activate the proto-oncogenes (Szyf 2005).

Increased regional methylation of CpG islands is evident within the promoter regions of many genes in cancer cells. This hypermethylation is associated with the inhibition of transcription and the silencing of tumor suppressor genes involved in tumor progression.

1.4 Restriction Enzymes

1.4.1 Sel1

The recognition sites for Sel1 are vCGCG, GCGCv. It is originated from Synechococcus elongatus bacteria.

1.4.2 BmgB1

The recognition sites for the restriction enzyme, BmgBI, are CACvGTC, GTGvCAG. It is originated from Bacillus megaterium. A 50 μL reaction containing 1 μg of DNA and 50units of BmgBI incubated for 16 hours at 37°C resulted in a DNA pattern free of detectable nuclease degradation as determined by agarose gel electrophoresis. The optimum temperature for BmgBI is 37oC and it has to be heat-inactivated for 20 minutes at 65oC.

1. 5 Tumor Suppressor Genes

Tumor-suppressor genes are critical in the regulation of cellular division. They encode proteins that inhibit the progression of cells during cell division. Unlike proto-oncogenes, these tumor suppressors are activated within the normal genome. The tumor suppressor gene can either stop cell division until damage is rectified when DNA damage is detected, or induce the cells to undergo programmed cell death (apoptosis). Should the tumor suppressor genes become impaired (achieved by CpG-island methylation), then cells will proliferate in an uncontrolled manner, creating more DNA damage. Inactivation of the tumor suppressor gene is closely associated with tissue specificity for cancer development (Esteller 2005).

(1) Retinoblastoma 1 Gene (RB1)

The official full name of RB1 is retinoblastoma 1. It is also known as RB; pRb; OSRC; pp110; p105-Rb; RB1 . It is a protein coding gene which is a negative regulator of the cell cycle. It is also the first tumor suppressive gene discovered.

RB1 encodes a nuclear phosphoprotein that is critical in regulating the cell cycle such that it helps stabilize the constitutive heterochromatin to keep and maintain the overall chromatin structure (Abouzeid, et al. 2009). The hypophosphorylated active form of this protein binds to the transcription factor E2F1. Defects in this gene due to the inactivation of both alleles in this gene can result in retinoblastoma in children (blindness from therapeutic eye ablation), bladder cancer and osteogenic sarcoma. Children who inherit one defective copy of the RB1 gene can have an increased chance of having retinoblastoma (Gonzalez-Gomez, et al. 2003). The RB1 gene can be inactivated by a combination of genetic and epigenetic alterations of two alleles. The RB1 gene contains a CpG island that house the critical promoter region, which is unmethylated during development (Donovan, et al. 2006; Richter, et al. 2003).

RB1 is located on the long arm of chromosome 13q14, chromosome 13: 48,877,911-49,056,122 forward strand.

There are 4 transcripts in this gene:

Table 2.1 Transcripts of RB1 gene


Transcript ID

Protein ID




No protein product








No protein product




No protein product


RB1 has 27 Exons, its transcript length is 4,840 bps and the translation product has 928 amino acid residues.

Analysis of 5'UTR and 1 st Exon of RB1 -002 (ENST00000267163)

Lower case (5'-UTR) and upper case (Exon 1)


Analysis of CpG island on RB1 Gene

Length of DNA:700

Parameters used to find CpG islands:

Minimum length of Island:300

Maximum length of Island:2000

C+Gs/Total bases > 50%

CpG observed/CpG expected > 0.6

CpG Islands found:1


Analysis of the 5'UTR and Exon 1 of RB1 gene also show that it has a number of CpG dinucleotides and two CGA within the first exon. Almost the entire length of the 700 bp 5'UTR-Exon1 sequence constitute a CpG island starting from first nucleotide to 694th nucleotide with a total of 79 CpG dinucleotides.

In our study, we divide the RB1-Exon 1 region into two parts as shown to analyse the entire length in two separate PCR reactions. 5'UTR-RB1 constitutes the first 268 nucleotides of the 700 bp sequence whilst RB1-Exon1 comprises the remaining 432 nucleotides. The two sections of the genes were then subject to the analysis by the Methyl Express Primer Design Software to obtain the best primer sets for methylated, unmethylated and bisulphite converted DNA. The sequences of the primers are tabulated as in Table 3.1.

Gggatgagaggtggggggcgccgcccaaggagggagagtggcgctcccgccgagggtgcactagccagatattccctgcggggcccgagagtcttccctatcagaccccgggatagggatgaggcccacagtcacccaccagactctttgtatagccccgttaagtgcaccccggcctggagggggtggttctgggtagaagcacgtccgggccgcgccggatgcctcctggaaggcgcctggacccacgccaggtttcccagtttaa ttcctcatgacttagcgtcccagcccgcgcaccgaccagcgccccagttccccacagacgccggcgggcccgggagcctcgcggacgtgacgccgcgggcggaagtgacgttttcccgcggttggacgcggcGCTCAGTTGCCGGGCGGGGGAGGGCGCGTCCGGTTTTTCTCAGGGGACGTTGAAATTATTTTTGTAACGGGAGTCGGGAGAGGACGGGGCGTGCCCCGACGTGCGCGCGCGTCGTCCTCCCCGGCGCTCCTCCACAGCTCGCTGGCTCCCGCCGCGGAAAGGCGTCATGCCGCCCAAAACCCCCCGAAAAACGGCCGCCACCGCCGCCGCTGCCGCCGCGGAACCCCCGGCACCGCCGCCGCCGCCCCCTCCTGAGGAGGACCCAGAGCAGGACAGCGGCCCGGAGGACCTGCCTCTCGT

(2) RUNX3

The official full name of RUNX3 is runt-related transcription factor 3. It is a RUNX family gene located on the distal portion of the short arm of human chromosome 1p36, 30.31 kb-sized, composed of 5 exons, which is frequently deleted. RUNX3 is related to the carcinogeneses of several solid tumors, including gastric cancer. It is also known as AML2; CBFA3; PEBP2aC; FLJ34510; MGC16070; RUNX3. It is also a protein coding gene which encodes a member of the runt domain-containing family of transcription factors. The heterodimer of this protein together with a beta subunit forms a complex that binds to the core DNA sequence 5'-PYGPYGGT-3', present in several enhancers and promoters which can activate or suppress transcription.

It can also interact with other transcription factors. Besides, RUNX3 can serve as a tumor suppressor whereby the gene is usually deleted or transcriptionally silenced in cancer. Several transcript variants encoding different isoforms have been found for this gene (Nomoto, et al. 2008).

There are many mechanisms, which include promoter region hypermethylation, loss of heterozygosity, hemizygous deletion and mutation, are associated with the downregulation of the RUNX3 gene, which leads to the carciongeneses of several human solid tumors (Hwang, et al. 2007). It is a RUNX family gene located on the distal portion of the short arm of human chromosome 1p36, 30.31 kb-sized, composed of 5 exons, which is frequently deleted.

Ensembl search shows that there are 7 transcripts in this gene.

Table 2.2 Transcripts of RUNX3 gene


Transcript ID

Protein ID












No protein product




No protein product














RUNX3-001 has 5 exons and the first exon (ENSE00001406046) consists of 693 nucleotides as shown below with the following sequence. It has a number of CG dinucleotide repeats as well as CGA trinucleotides as shown in red and underlined respectively.

>RUNX-3 Exon1


Analysis of the gene for CpG islands of RUNX-3-001 Exon1

Length of DNA:693

Parameters used to find CpG islands:

Minimum length of Island:300

Maximum length of Island:2000

C+Gs/Total bases > 50%

CpG observed/CpG expected > 0.6

CpG Islands found:1

Start: 1 End: 685 Length:685

A%: 10.51 -- T%: 11.09 -- C%:38.98 -- G%:39.42

C+G%: 78.39 -- CpG%: 15.50



Almost the entire length of the exon 1 sequence constitute a CpG island starting from first nucleotide to 685th nucleotide with a total of 106 CpG dinucleotides.

The CpG island of the RUNX-3 exon 1 sequence was subjected to the Methyl Express primer design software tool to find out the best primer pairs for methylated, unmethylated and bisulphite converted DNA. The sequences of the primers are as listed in Table ???.

1.6 Methodologies for DNA Methylation Studies

The conversion of cytosine to 5-methylcytosine is a critical epigenetic change in our genome. This methylation is associated with gene silencing whereby genes rich in 5-methylcytosine content are normally transcriptionally silent. Thus methods used to study DNA methylation are based on this theory (Coleman and Tsongalis 2006).

Methylation-sensitive Restriction Endonuclease Analysis

Methylation-sensitive restriction enzymes are usually used with Southern blot analysis to compare the resultant DNA restriction fragments, thereby determinating the potential methylated sites of a selected genome.

Genomic DNA is cleaved into fragments by methylation-specific restriction enzyme and subjected to heat inactivation to terminate the enzymatic action. PCR is then performed with specific primers flanking the digestion sites thereby allowing the amplification of the DNA sequence of interest. The amplicons are then separated by gel electrophoresis and the digestion pattern is analysed.

However, this method requires large amounts of DNA and it can detect methylation provided that more than a few percent of alleles are methylated. The methylated sites can only be analysed if it is located within the recognition site of the restriction enzyme and also only if they are fully are completely methylated (in both strands) or completely unmethylated. This method cannot be used to analyse hemimethylated sites. In additon, a false-positive result may arise due to the incomplete restriction digestion of the DNA.

Methylated DNA Immunoprecipitation (MeDIP)

MeDIP employs immunoprecipitation approach using antibody specific for methylated cytosine to precipitate DNA that is methylated. The genomic DNA is first broken down to fragments by means of sonication then precipitated with a monoclonal antibody (anti-5-methycytidine (5mC)). The resultant enriched MeDIP fragments can be analysed by either PCR or Microarray analysis. From the anaylsis, low enrichment intensity can indicate an unmethylated status or the absence of CpGs in the selected sequence.

MeDIP relies on both the methylation state of the target sequence and the amount of CpGs it house within. Large quantity of genomic DNA will then be required for the analysis that can be labor intensive and time-consuming (Weber and Schübeler 2006).

Bisulfite Conversion

Bisulfite conversion consists of subjecting genomic DNA to sodium bisulfite-induced oxidative deamination under conditions whereby cytosine is converted to uracil whereas 5-mC remains unconverted. The use of bisulfite treatment prior to PCR allows the methylated cytosine in the genomic DNA to be directly identified. Due to this bisulfite treatment, the DNA strands are no longer complementary. The target sequence is then amplified by PCR with strand specific primers to differentiate between the methylated and unmethylated sequences. However, if the bisulfite conversion is incomplete or insufficient, then the accurate analysis of the methylation profile cannot be done. Furthermore, this technique would require the genomic DNA to be fully denatured prior to bisulfite treatment.

Direct Sequencing

The PCR products from bisulfite conversion can be used directly for sequencing which indicates the average estimated density of methylation and allows primers to be designed successfully given the methylation propensity of each CpG sites within the selected sequence. However, if the methylation levels are low (less than 25% at any one site), then direct sequencing will not be sensitive enough to produce reliable methylation detection.

Methylation-specific PCR (MSP)

MSP-PCR is the most commonly used assay for detecting methylation. MSP is based on the theory that primers with a mismatched 3' residue cannot function (anneal). MSP is positive when the 3' end of each primer recognizes the methylated cytosine residues on both strands. After bisulfite conversion of DNA, PCR is carried out using primers that are specific for either methylated or unmethylated DNA. Amplification of either the methylated or unmethylated DNA will occur prior to successful annealing of the primer. Gel electrophoresis is then used to resolve the amplified products. Band formation indicates that the sequence is methylated or unmethylated depending on the primers-specificity used. Hence MSP is an ideal technique to screen CpG islands that are heavily methylated.

However, MSP is susceptible to false positive results due to false priming even though methylation is absent in the target sequence or low methylation levels in the target sequence. Furthermore amplification can happen across the 3' mismatch region if the annealing temperature is too low.

Methylation-sensitive Single-Nucleotide primer Extension assay (MS-SNuPE)

This assay is used to determine the quantitative levels of cytosine and thymine at a single CpG site. The bisulfie converted DNA is amplified by PCR, then hybridized with a primer that has a sequence that is adjacent to the cytosine of a CpG residue. This primer can only extend with the appropriate deoxynucleotide triphospate, deoxycytidine triphosphate for methylated cytosine and thymidine triphospate for unmethylated cytosine. The PCR product is resolved by gel electrophoresis. MS-SNuPE assay is quantitative but is too laborious for screening applications despite its capability to analyse several CpG sites in a single reaction with different primers.

Combined Bisulphite Restriction Analysis (COBRA)

This approach can be used to estimate the level of methylation of DNA. The amplified bisulfite converted sequence may contain several restriction enzyme cutting sites. These restriction enzymes can recognize methylated or unmethylated DNA. One such restriction enzyme is BstU1. Its recognition site CGCG is resistant to bisulfite conversion when it is methylated, but becomes TGTG when it is unmethylated. Restriction digestion is used together with methylation independent primers (do not have CpG dinucleotides) that can amplify both the methylated and unmethylated DNA. After amplification, the DNA is digested and the digested fragment is resolved by gel electrophoresis. This technique is also known as Combined Bisulfite modification Restriction Analysis (COBRA). However, as many CpG sites cannot be analysed, this approach cannot detect the methylation restricted to the untested regions. Furthermore, this analysis requires the complete conversion of the DNA.

Illumina Methylation Assay

Modern methods for genomic DNA methylation studies include using of microarrays or bisulfite- genomic DNA sequencing of chosen regions which are very effective. However these methods are rather labor intensive, which require large amounts of DNA sample and rely heavily on bioinformatic analyses, resulting in difficulties in large-scale studies where sample quantity may be limited.

In this assay, quantitative measurements of DNA methylation are determined for 27578 CpG dinucleotides across 14,495 genes. Bisulfied converted DNA is annealed to sequence specific primers linked to two individual bead types corresponding to each CpG locus, one to the methylated (C) and the other to the unmethylated (T) locus. This hybridization is followed by single base extension with DNP and Biotin-labelled ddNTPs. After extension, the array is fluorescently stained and the intensities of the bead types are measured which corresponds to the quantity of unmethylated and methylated sequences present in the genomic DNA.

For this assay, only 1µg of bisulfie converted genomic DNA is required. Furthermore, it has the ability to provide genome-wide coverage for 12 samples simultaneously and it is easily capable of high-throughput analyses. However, this method is not cost effective for small-scale studies.

Methylation-Specific Oligonucleotide (MSO) Microarray

It is a gene-specific hybridization approach used to screen methylation across several CpG sites of numerous genes simultaneously. The bisulfite converted DNA is amplified with primers that do not contain CpG dinucleotides, pooled, purified and labeled with fluorescent dye then hybridized to arrayed oligonucleotide probes of both methylated and unmethylated regions. DNA target binding to the probe (intensity) is dependent on the amount of methylated and unmethylated regions in the target sequence.

MSO has the capability to differentiate even single-nucleotide differences within CpG sites, not just sequences where motifs are restricted as in restriction digestion approaches. Hence large amount of CpG sites can be specifically probed by this approach, which requires very little quantity of DNA. However, this approach is relatively expensive and there is the possibility of cross-hybridization between the mismatch probes and targets.

Real-Time PCR and High Resolution Melt Curve Analysis

It is a method used to monitor the progress the PCR reaction in real time. It can also quantify PCR product concurrently. Real-time PCR is based on the detection of fluorescence emitted by a DNA binding dye, which increases as the product accumulates with each successive amplification cycle. The intensity of fluorescence at each cycle can help monitor PCR reaction during the exponential stage.

Bisulfite converted DNA is amplified by real-time quantitative PCR, using strand specific PCR primers flanking an oligonucleotide probe with a fluorescence DNA binding dye (SYBR-Green- 1). SYBR-Green binds to the minor groove of any double stranded DNA.

In the PCR reaction solution, the unbound dye emits very little fluorescence. The fluorescence is enhanced upon binding of the dye to the double stranded DNA. Furthermore, the primers used are sequence specific, only methylated DNA (double-stranded) will be amplified and thus emits fluorescence. As more double stranded PCR products (amplicons) are produced, the intensity of SYBR-Green dye signal increases. Therein indicates higher levels of methylation in the target sequence.

Although real-time PCR assay using SYBR-Green 1 is cheap and easy to use, the main setback to intercalation based detection of PCR product accumulation is signal generation from both specific and non-specific amplicons.

High-resolution melting curve analysis (HRM) is a simple, cost effective closed-tube fluorescent technique that can be used for the detection of aberrant DNA methylation by exploiting the differences in melting temperature, Tm, between the methylated and unmethylated alleles following bisulfite conversion.

Traditional methylation detection techniques employs the two-step approach, consisting of an initial PCR amplification followed with the subsequent product analysis usually by gel electrophoresis. HRM is a more advantageous approach, as it is rapid and secure due to its closed tube properties (in-tube PCR assay), which omits the laborious gel separation step whilst reducing any potential risk of contamination. Furthermore, with the exception of genomic sequencing, traditional methods are limited to the analysis of one or a few CpG sites in each setting. HRM is also easy to use as it does not require any post PCR sample manipulation other than the closed tube melting analysis. Automation of HRM is also possible as very few steps are involved. HRM is a non-destructive technique whereby each sample can be re-analysed and sequenced without fear of sample degradation. HRM is also relatively cheap as it requires only PCR reagents and fluorescent dye. Lastly, HRM is highly suited for high-throughput mutation screening the HRM instruments allows the simultaneous acquisition of up to 384 fluorescent melting signals under five minutes (Isaacs and Rebbeck 2007).

Double stranded DNA melts in a series of steps when gradually heated, whereby each step represents the melting of a particular discrete segment or melting domain. Tm of the melting domain increases with the increase in GC content. After bisulfite conversion, the methylated and unmethylated sequences can then be differentiated due to their differences in GC content and in turn, their thermal stability (Bruns, et al. 2007).


1. Hana Abouzeid,Daniel F. Schorderet,Aubin Balmer,and Francis L. Munier , 2009. Germline mutations in retinoma patients: Relevance to low-penetrance and low-expressivity molecular basis [online]. Available from: http://www.ncbi.nlm.nih.gov:80/pmc/articles/PMC2671583/?tool=pmcentrez [Accessed 12 November 2009]

1. P Gonzalez-Gomez, M J Bello,M E Alonso,D Arjona,J Lomas,J M de Campos,A Isla,and J A Rey , 2003. CpG island methylation status and mutation analysis of theRB1gene essential promoter region and protein-binding pocket domain in nervous system tumours [online]. Available from: http://www.ncbi.nlm.nih.gov:80/pmc/articles/PMC2376780/?tool=pmcentrez [Accessed 12 November 2009]

3. Stacy L Donovan,Brett Schweers,Rodrigo Martins,Dianna Johnson,and Michael A Dyer, 2006. Compensation by tumor suppressor genes during retinal development in mice and humans [online]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1481602&tool=pmcentrez [Accessed 12 November 2009]

4. Suzanne Richter,Kirk Vandezande,Ning Chen,Katherine Zhang,Joanne Sutherland, Julie Anderson,Liping Han,Rachel Panton, Patricia Branco,and Brenda Gallie, 2003. Sensitive and Efficient Detection ofRB1Gene Mutations Enhances Care for Families with Retinoblastoma [online]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=379221&tool=pmcentrez [Accessed 12 November 2009]

5. José R Valverde,Javier Alonso,Itziar Palacios,and Ángel Pestaña, 2005. RB1gene mutation up-date, a meta-analysis based on 932 reported mutations available in a searchable database [online]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1298292&tool=pmcentrez [Accessed 12 November 2009]

6. Jason I Herschkowitz,Xiaping He,Cheng Fan, and Charles M Perou, 2008. The functional loss of the retinoblastoma tumour suppressor is a common event in basal-like and luminal B breast carcinomas [online]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2614508&tool=pmcentrez [Accessed 13 November 2009]

7. Erika M. Wolff,Ganging Liang,Connie C. Cortez,Yvonne C. Tsai,J. Esteban Castelao,Victoria K. Cortessis,Denice D. Tsao-Wei,Susan Groshen,and Peter A. Jones, 2009. RUNX3Methylation reveals that Bladder Tumors are Older in Patients with a History of Smoking [online]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2536768&tool=pmcentrez [Accessed 13 November 2009]

8. S Nomoto,T Kinoshita,T Mori, K Kato,H Sugimoto,N Kanazumi,S Takeda,and A Nakao, 2008. Adverse prognosis of epigenetic inactivation inRUNX3gene at 1p36 in human pancreatic cancer [online]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2391125&tool=pmcentrez [Accessed 13 November 2009]

9. Ki-Tae Hwang,Wonshik Han,Ji-Yeon Bae,Sung Eun Hwang,Hyuk Jai Shin,Jeong Eon Lee,Sung-Won Kim,Hyun Jung Min,and Dong-Young Noh, 2007. Downregulation of theRUNX3Gene by Promoter Hypermethylation and Hemizygous Deletion in Breast Cancer [online]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2694388&tool=pmcentrez [Accessed 13 November 2009]

10. RUNX3runt-related transcription factor 3 [Homo sapiens] [online]. Available from: http://www.ncbi.nlm.nih.gov/gene/864?ordinalpos=2&itool=EntrezSystem2.PEntrez.Gene.Gene_ResultsPanel.Gene_RVDocSum [Accessed 13 November 2009]

11. Real time PCR, Premier Biosoft international. [online]. Available from: http://www.premierbiosoft.com/tech_notes/real_time_PCR.html [Accessed 13 November 2009]

12. Marcel W. Coolen, Aaron L. Statham, Margaret Gardiner-Garden, and Susan J. Clark, 2007. Genomic profiling of CpG methylation and allelic specificity using quantitative high-throughput mass spectrometry: critical evaluation and improvements [online]. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2094090&tool=pmcentrez [Accessed 13 November 2009]

13. Sel1 restriction enzyme, REBASE. [online]. Available from: http://rebase.neb.com/rebase/enz/SelI.html [Accessed 13 November 2009]

14. Real time PCR: the TaqMan Method. [online]. Available from: http://www.bio.davidson.edu/Courses/Molbio/MolStudents/spring2003/Pierce/realtimepcr.htm [Accessed 13 November 2009]

15. Michaël Weber and Dirk Schübeler. Methylated DNA Immunoprecipitation (MeDIP) [online]. Available from: http://www.epigenome-noe.net/researchtools/protocol.php?protid=33 [Accessed 13 November 2009]

16. DNA methylation and CpG island. [online]. Available from http://www.web-books.com/MoBio/Free/Ch7F2.htm [Accessed 14 November 2009]

1. Margaret A. Knowles, Peter Selby, 2005. Introduction to the cellular and molecular biology of cancer. 4th Edition. Oxford University Press.

1. Trygve O. Tollefsbol, 2004. Epigenetics Protocols. Humana Press.

3. Andrew Wallace Hayes, 2007. Principles and methods of toxicology. CRC Press.

4. Moshe Szyf, 2005. DNA methylation and cancer therapy. Springer.

5. Attila T. Lorincz, 2006. Nuclaic acid testing for human disease. CRC Press.

6. Eberhard Passarge, 2007. Color atlas of genetics. 3rd edition, revised. Thieme.

7. William B. Coleman, Gregory J. Tsongalis, 2006. Molecular diagnostics: for the clinical laboratorian. 2nd editon. Humana Press.

8. Manel Esteller, 2005. DNA methylation: approaches, methods and applications. CRC press.

9. Ken I. Mills, Bernie H. Ramsahoye, 2002. DNA methylation protocols. Humana press.

10. Erika Kress-Rogers, Christopher J. B. Brimelow, 2001. Instrumentation and sensors for the food industry. 2nd edition. Woodhead Publishing.

11. Clausdine Isaacs, Timothy R. Rebbeck, 2007. Heriditary Breast Cancer. CRC Press.

12. Davis E. Bruns, Y. M. Dennis Lo, Carl T. Wittwer, 2002. Molecular testing in laboratory medicine: Selections from Clinical Chemistry, 1998-2001, With annotations and updates. Reprinted edition. Amer. Assoc. for Clinical Chemistry.

13. David E. Bruns, Edward R. Ashwood, Carl A. Burtis, 2007. Fundamentals of molecular diagnostics. Elsevier Health Sciences