DNA methylation and transcription

== DNA methylation affects TF binding and transcription ==

Introduction: It's commonly accepted that DNA methylation of a promoter repress transcription of this gene in normal tissues. Recently, a class of actively expressed genes having relatively methylated promoters has been discovered.

Purpose:

• To explore the idea that DNA methylation affects TFBS, preventing some TF from binding to DNA, and therefore represses transcription.
  
• To select TFs most likely sensitive to DNA methylation (first, in human, possibly in other species)
  

We plan to do:

1. Functional analysis
- Predict TFBS (homotypic clusters, composite elements?, conserved sites?) in promoters
- Estimate methylation level of each TFBS (may be, only core positions)
- Calculate correlation between level of methylation and expression from given promoter for each TFBS
- Select promoters and corresponding TFs with highest and lowest cc

2. Evolutionary analysis
- Predict TFBS (homotypic clusters, composite elements?) in promoters
- For whole set of given TF's binding sites estimate probability of C>T SNP (in CG or in CNG) and of C>T interlineage substitution

We have data for functional analysis:

Primary cells:
peripheral blood mononuclear cells PBMC (genome-wide BS-seq + CAGE)

Andreas Lennartsson
I can provide single bp resolution CpG methylation data from Illuminas Infinum 27k and 45k bead array, for the hematopoietic cell populations (common myeloid progenitor, granulocyte-monocyte progenitor, myelocyte-promyelocyte and band cells/neutrophils)+CAGE. In addition, I have CpG methylation data from almost 150 acute myeloid leukemia (AML) patient samples from the same platforms + Illumina expression array data for 15 of the AML patients.

Tissues:
frontal cortex (genome-wide MeDIP-seq) - frontal lobe (CAGE)

We are looking for data: DNA methylation with single bp resolution (genome-wide or covering not less then 1% of genome) in primary cells, tissues or cell lines represented in FANTOM5

Collaboration: Piotr (you are invited to add your plans of study on influence of TF on methylation)

Sung-Joon Park (linear regression model for gene expression and methylation)
- Absolute gene expression is explained by summing up the "effects" of TFBS activities and DNA methylation levels.
- The first step to do is to identify functional "effects".
- This will be done by a feature selection procedure based on model reduction or some statistical tests.
- The feature selection will suggest which TFBS and/or methylation are indispensable to explain the gene expression.

== TF binding affects DNA methylation ==

Collaborators are very welcome. If you have any ideas how to improve the research, please contact me directly or add your suggestion here.

Martin Frith
By coincidence, I am working (outside Fantom) on methods to align bisulfite-converted DNA accurately and efficiently. This is very much a technical detail, but if you're interested, I'd be glad to share what I have.