no DNA fragment pull down in the first CpG island (MBD-seq) - (Sep/13/2012 )

Hallo everyone,

I performed a methyl-CpG binding doman (MBD) protein capture and genome-wide sequencing.
To see any methylation level changes in those gene I am interest in.
After I got the data, using IGV to see the reads, I found that most of the gene's first CpG island (start from TSS) has almost no reads!

Same situation in this journal:
Dnmt3a-Dependent Nonprmoter DNA Methylation Facilitates Transcription of Neurogenic Genes Science 329,444 (2010) fig2D

My question is:
1) Suppose MBD will capture the methylated DNA fragments, why most of the genes have no reads in the first CpG island?
2) Why only happen in the first CpG island of the genes but not the second or third even last one?
3) Is that any method can come over this problem? ( If this situation is a problem or bias)

Thanks in advance!

kb

What genes are you looking at--expressed or non-expressed genes? The result may not necessarily wrong. Normally CpG island in the promoters of actively transcribed genes are not methylated, only some genes are abhorrently methylated and thus silenced in transcription. If you do ChIP-seq using mark such as H3K4me3, a mark for active transcription, you would find a good portion of genes have the mark on their TSS, these genes would certainly have no MBD signal at their TSS.

The cells for MBD-seq is a stable cell line which overexpress an oncogene. I am just looking at any genes will change their methylation level under this circumstances.
That means if the proximal promoter CpG island have methylation level changes, those genes I have to pay attention on it as it was activated by the oncogene?