Low reads mapping in methylation sequencing

Question

[SJ Basu]

Hello People,

I have four methylation samples (MeDIP-seq) of two related plant species, testing methylation in case of saline stress. details are as follows:

Sample 1: Species 1 no stress

Sample 2: Species 1 with saline stress

Sample 3: Species 2 no stress

Sample 4: Species 2 with saline stress

Now when I map these reads to the genome (using BWA) I get 95 % mapping for Sample 1 and 3 while Sample 2 and 4 shows 23 % mapping. I am not able to understand the reason. All samples have ~25 mill reads and the sample extraction and quality of reads really is good !! Has anyone faced this problem ?? What can be the cause ??

Answer 1

[Chris Miller]

Pull them up in IGV - did you get read-through into the adapters? Then adapter trimming might help. You might also try using a methylation aware aligner (BSmap, Biscuit, etc)/

[SJ Basu]

Hello Chris,

I followed your advice and used Bismark (in Bowtie1 mode) and the mapping percentage dropped even more at 3%.

Is it that the methylated-Cytosines (C) are converted to thymine (T) in the methylated DNA extracted and hence the sequenced read is also having T whereas in reference genome all these points with T are actually C ??!!

Any idea what can be going wrong ??!!!