0
$\begingroup$

What is the role of DNA methylation in breast cancer? DNA methylation is a process by which methyl groups are added to the DNA molecule.

In the September 2015 paper The Role of Methylation in Breast Cancer Susceptibility and Treatment published in Anticancer Research, Pouliot, Labrie, Diorio and Durocher write:

DNA methylation is a critical mechanism of epigenetic modification involved in gene expression programming, that can promote the development of several cancers, including breast cancer. The methylation of CpG islands by DNA methyltransferases is reversible and has been shown to modify the transcriptional activity of key proliferation genes or transcription factors involved in suppression or promotion of cell growth. Indeed, aberrant methylation found in gene promoters is a hallmark of cancer that could be used as non-intrusive biomarker in body fluids such as blood and plasma for early detection of breast cancer.

In addition, Pouliot, Labrie, Diorio and Durocher state that:

DNA methylation is a reversible mechanism that occurs most commonly with an addition of a methyl group in the fifth position of the pyrimidine ring of cytosine on CpG sites within the genome. Human DNA methylation is introduced into the sequence of nucleotides by enzymes of the DNA cytosine methyltransferases family including DNMT1, DNMT3A, DNMT3B and DNMT3L (6).

The DNMT1 protein methylates newly-synthesized strands prior to chromatin packaging and is localized to the replication foci during the S-phase. This highly expressed enzyme is mainly responsible for maintenance of the methylation pattern during replication (7). An increased expression of DNMT1 is observed in many types of cancers and this overexpression is associated with cellular transformation, while reduced DNMT1 expression levels seem to be associated with a protective effect (8, 9).

The DNMT3A and -B proteins allow de novo DNA methylation activity in vitro, without distinction between unmethylated and hemi-methylated DNA (10). In particular, the expression of DNMT3B is elevated in several human cancer types, while its suppression results in tumor cell apoptosis (11, 12). Indeed, inactivation of the TSG Ras association domain family 1 isoform A (RASSF1A), through promoter hypermethylation triggered by up-regulation of DNMT3B, is a common event in numerous cancers or tumor types (13). Several studies suggest that DNMT3B plays a predominant role over DNMT3A and DNMT1 in breast tumorigenesis, given its overexpression in breast cancer tissues compared to DNMT1 and -3A (14). Although DNMT3L has no catalytic activity on its own, this enzyme promotes the activity of DNMT3A and -B by increasing their binding capacity for the methyl group donor, S-adenosyl-L-methionine (15).

In mammalian tissues, DNA methylation occurs in CpG dinucleotides, and clustering of these elements in the 5’ regulatory regions of approximately 60% of genes are referred to as CpG islands (16). Unusual methylation of these CpG islands may lead to silencing of certain genes involved in key proliferation and apoptosis pathways such as TSGs, DNA repair and hormone receptor genes, as well as genes that inhibit angiogenesis (17). Two different DNA methylation mechanisms lead to transcriptional gene repression: CpG dinucleotide methylation inhibits the binding of transcriptional factors to their promoter regulatory elements; and methylated cytosines enhance the recruitment of methylated binding domain proteins, which inhibit binding of transcription factors through inactive chromatin configuration state around the genes (18). On the other hand, ten-eleven translocation (TET) proteins such as TET2 can remove DNA methyl marks and trigger re-expression of silenced genes. Of interest, TET2 has been shown to be mutated and therefore inactivated in many types of cancer (19).

In addition, growing evidence suggests that methylation of regions located upstream of CpG islands, called CpG island shores, as well as intragenic sequences, are also important for regulation of gene expression and can be involved in disease development and progression (20, 21). Generally in normal cells, gene promoters are unmethylated, while gene bodies and intergenic regions are methylated.

How do you analyze DNA methylation data generated from the EPIC array from Illumina as described here,  https://www.illumina.com/products/by-type/microarray-kits/infinium-methylation-epic.html?

Analysis of DNA methylation can be assessed through different methods including bisulfite conversion. This can be performed in combination with real-time methylation-specific PCR and specific primers to amplify CpG islands of a panel of selected genes (2). A genome-wide approach with high-throughput sequencing can also be performed using bisulfite-converted DNA from different non-invasive biological sources such as whole blood, serum and plasma.

The rapidGSEA method described in https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-016-1244-x takes five default arguments and three optional arguments rapidGSEA <- function(exprsData, labelList, geneSets, numPermutations, metricString, dumpFileName="", checkInput=TRUE, doublePrecision=FALSE) {...} metricString denotes the local ranking measure (one of the following) and numPermutations denotes the number of permutations in the resampling test.

If metricString denotes the local ranking measure we are requesting that rapidGSEA use when it is invoked, which metricString value listed below should I choose so that rapidGSEA properly analyzes DNA methylation data generated from the EPIC array from Illumina ?

naive_diff_of_classes
naive_ratio_of_classes
naive_log2_ratio_of_classes
stable_diff_of_classes
stable_ratio_of_classes
stable_log2_ratio_of_classes
onepass_signal2noise
onepass_t_test
twopass_signal2noise
twopass_t_test
stable_signal2noise
stable_t_test
overkill_signal2noise
overkill_t_test

I finally found the answer to the question, " If metricString denotes the local ranking measure we are requesting that rapidGSEA use when it is invoked, which metricString value should I choose so that rapidGSEA operates in a manner most closely resembling fgseaL?"

Please read this great article in May 12 2017 BioMed Central (BMC) Bioinformatics article titled with the URL, https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-017-1674-0

Also, please read this blog , "Diving into Genetics and Genomics: Gene Set Enrichment Analysis (GSEA) explained." with the URL http://crazyhottommy.blogspot.com/2016/08/gene-set-enrichment-analysis-gsea.html

After reading these two articles, my choice for the best rapidGSEA local ranking measure is Minimum Significant Difference(i.e. MSD) because it has the best overall false positive rate(i.e FPR) for larger sample sizes.

Finally, it is important to realize that fgseaL's phenotype labeling principle cannot be emulated by GSEA or rapidGSEA with any of it's sixteen possible ranking metrics.

Thank you.

$\endgroup$

closed as unclear what you're asking by anongoodnurse, canadianer, David, kmm, Bryan Krause Aug 21 '17 at 14:42

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    $\begingroup$ Can you please use > at the start of your quotations and link to the primary sources (not every cited papers in the quotation, but at list the source you quote)? $\endgroup$ – Remi.b Aug 17 '17 at 23:22
  • $\begingroup$ @Remi.b The source I quote from is the September 2015 paper "The Role of Methylation in Breast Cancer Susceptibility and Treatment" published in Anticancer Research by Pouliot, Labrie, Diorio and Durocher. $\endgroup$ – Frank Aug 18 '17 at 6:56
  • $\begingroup$ @canadianer, Thank you for the very nice edit just now. $\endgroup$ – Frank Aug 18 '17 at 8:05
  • 1
    $\begingroup$ I'm voting to close this question as off-topic because this is not a question it is a review inviting a discussion. If you cannot answer your own question after so much reading you can hardly expect a simple definitive answer here. $\endgroup$ – David Aug 18 '17 at 20:36
  • $\begingroup$ @David,I just found this article github.com/gravitino/cudaGSEA/blob/master/cudaGSEA/man/GSEA.Rd which answers my question,Thank you $\endgroup$ – Frank Aug 18 '17 at 22:35
2
$\begingroup$

Okay I'll try to put together a quick answer.

First: 'How do you analyze DNA methylation data generated from the EPIC array from Illumina'

The most common method to analyse DNA methylation is based on bisulfite conversion of DNA. This chemical converts the base cytosine (C) to uracil (U), which 'reads' the same as thymine (T), but it only works on non-methylated cytosine. When you want to know which bases in your DNA are methylated, you split your sample in two parts, do the bisulfite conversion on one and then sequence both. The C/T differences in the sequence then tell you, which C's in your sequence were methylated. The Illumina arrays are special sequencing methods, that do not look for all sequences in a given sample, but only check specific ones (usually some genes of interest). For methylation analysis these chips allow the sequencing of the two versions (normal & bisulfite converted) of each region of interest separately.

Now: 'Can DNA methylation induce breast cancer?'

In principle yes, but it is very unlikely/almost impossible to be the only factor. A cancer forms only, if a whole range of cellular functions (esp. growth) are abnormally regulated and if the various control mechanisms intended to prevent this fail. Therefore most cancers have hundreds to thousands of mutations and some of these might also be abnormal DNA methylation patterns or regulation. Further details of how DNA methylation can affect which genes are active or repressed is also discussed in these questions.

$\endgroup$
  • $\begingroup$ Could you please select the rapidGSEA * gene set enrichment analysis * metric string value shown at the bottom of the original question which most properly measures the statistical significance of rejecting the null hypothesis where we are comparing a normal healthy control group of subjects with a group of subjects diagnosed with a distinct category of breast cancer? Thank you. $\endgroup$ – Frank Aug 18 '17 at 19:33

Not the answer you're looking for? Browse other questions tagged or ask your own question.