Number of Cs reported in ATCGmap

[akramdi]

Hi,

I am working with a reference genome that contains about 4310^6 Cs genome wide (tair10). I noticed that the reported number of cytosines in the ATCGmap output file does not match the number of reference Cs (I counted the number of Cs and Gs positions in ATCGmap file and found about 4210^6). Moreover, this total number of Cs reported is not constant between samples.

I assumed that the ATCGmap file would contain information for every reference C. Why this difference?

Best, Amira

[guoweilong]

Thanks for your question.

The ATCGmap file will only contain C site with at least 1 read covered.

Best, Weilong Guo

[akramdi]

Thanks for the quick reply!

Indeed, all reported Cs are covered by at least 1 read. I actually wondered about this possibility, so I checked the coverage information in the columns used to compute the methylation level (col $8 and 7$ if C in + strand ; col $14 and $11 if C in - strand), there were still Cs with 0 coverage ($8+$7=0 or $11+$14=0), so I ignored it.

Which brings me to another question about the methylation level (it might be trivial, please bear with me):

If we have a C on the + strand, why not consider $13 (# of reads from Crick strand mapped here, support C on Watson strand and G on Crick strand) as reads supporting a state of methylation?

Best, Amira

[guoweilong]

For a C on the Watson strand in reference genome, the $13 column would not contribute any information for the methylation status.

Actually, if you read the CGmap file, it would have fewer C/G sites, which only report positions those methylation-effective coverage (MEC) >=1.

Best, Weilong

[akramdi]

Got it, thanks!

Best, Amira