How to calculate the percent of methylated cytosines.

Question

[t2g4free]

I read from a paper:

"14% of cytosines are methylated in Arabidopsis thaliana, 4% in Mus musculus, 2.3% in Escherichia coli, 0.03% in Drosophila, and virtually none (< 0.0002%) in yeast"

but they did not say how to get the number of methylated cytosines. Anybody could show me how to calculate the percentage precisely?

[t2g4free]

Of course, they all use WGBS, after alignment, they got the cytosine methylation state in every reads, but then?

Answer 1

[Phil Ewels]

Tools such as Bismark and bwa-meth align WGBS reads and then extract the methylation status of every cytosine within each aligned read. They also produce summary statistics, including the genome-wide average methylation state.

For an example of what this looks like, you can see the Bismark documentation and an example summary report which reports the genome-wide methylation level.

[t2g4free]

Yeah, but there is no result just like the I read from the paper.

I know it must have to calculate, but how is the key point. Anyway, thx

[Phil Ewels]

If you want for all Cytosines, not context specific (CpG / CHG / CHH as printed in the report) just add the numbers:

• Total C's analysed 1218062196
• Methylated C's in CpG context 31593318
• Methylated C's in CHG context 1525159
• Methylated C's in CHH context 3319601

( (31593318 + 1525159 + 3319601) / 1218062196 ) * 100 = 2.9914792627 %

[dariober]

Maybe the OP is confused because "% methylated cytosines" could be interpreted as "% of genomic positions with a cytosine with some evidence of methylation". So if there are 10 cytosine positions in the genome and each of them has % methylation > 0, then "% methylated cytosines" = 100%. If this is what the OP (and the paper) means, then it's a bit trickier to calculate as one needs to define a threshold to claim that the C to T conversion at a genomic position is significantly smaller than 100% (i.e. methylation is greater than 0%) given sequencing and conversion errors.

[t2g4free]

btw, what does “OP” mean?

[GenoMax]

Original Post(er).

[t2g4free]

Thx 4 all your reply, @dariober's answer is what I mean.

I don't know what the paper means and confused.

[Phil Ewels]

If you post a link to the paper that would help to clarify the question :)