overlap judgment in bismark_methylation_extractor

[bio-Tao]

R1 ------>
R2 <------
Corresponding genomic coordinates:
start_R1 < start_R2
Example:
R1 : 100 ------ 250
R2 : 180 <------ 330
In this case:
end_R1 >= start_R2
250 >= 180
overlap = 250-180 = 70

R1 ------>
R2 <------
Example coordinates:
R2 : 100 <------ 250
R1 : 150 ------> 300
In this case:
end_R1 >= start_R2
300 >= 100
This condition is technically true, but the calculated overlap length:
overlap = 300-100 = 200
As a result, the trimming of the R2 methylation string may be incorrect.

Example：

but LH00524:324:22W7YFLT4:4:2407:14745:7404 ： incorrect calculation of overlap led to the loss of this read.

[FelixKrueger]

Could you please provide both R1 and R2, as well as the commands you used for alignment/extraction. I find it exceedingly difficult to get my head around all the strands and combinations this morning. Thanks

[FelixKrueger]

I have to say that I am not exactly sure what the screenshots provided are supposed to illustrate.

The first one appears to a cytosine report showing that pos 10058 has a coverage of 5 reads, all unmethylated.

The second one is ? Maybe the start of a coordinate sorted BAM file? I suppose numbers 1-6 are supposed to indicate that there are 6 reads (the labels are offset) over this region, but the coverage in the file above shows only 5 reads? The simplest explanation would probably be that 5 alignments are informative for the top strand, and 1 goes in the opposite direction and this reads methylation information from the opposite strand?

I don't want to fully exclude that something is fishy, but I think the easiest way to get to the bottom of this is to extract out all 6 reads from your FastQ files, and aligning just these 6 and then look at them it greater detail.

Also, just as a word of warning: This region appears to be very similar up to telomeric hexamer repeats which isn't uncommon for the very beginning of a chromosome (I just assume it starts with 10000 Ns). For telomeric repeats, the strand assignment may be incorrect, as read pairs may align to the top and bottom strands equally well in which case the top strand alignment is favoured. My initial feeling is that there may be an explanation for what is going on, but chances are that the results for this repeat region but not be trustworthy in any case. Just imagine the sequence was TTAGGGTTAGGGTTAGGG (the telomeric repeats), you would not get any methylation calls at all (as the sequence has no Cs....).

If you can provide the reads and make a convincing case I'd be happy to look in more detail.