首先OTOBCTOTCTOB都是描述测序reads的而不是描述参考基因组的。bisul-fate建库会将DNA双链文库中非甲基化的C转化成U。转化结束后被转化的U和互补链的G并不配对。此时正链OToriginal top strand和反链-OBoriginal bottom strand中均为C to T转换的原始reads。上述的reads经过PCR扩增后正反单链均产生完全互补链。OT的互补链为CTOTComplementary original top strandOB的互补链为CTOBComplementary bottom strand。可知CTOT和CTOB均为GA转换。图片引用自https://zhuanlan.zhihu.com/p/163495878正链是指FASTA 文件中提供的原始序列本身即参考基因组中记录的那条链。负链-是指FASTA 文件中记录的序列的反义链也就是将参考序列取反向互补得到的链。关于bismark比对可以参考这篇https://www.zxzyl.com/archives/759/比对生成的bam文件中XR字段如果是CT表示该reads是经过CT变换后匹配到了基因组中即该reads属于OB或者OT此时若XG字段为GA表示它mapping到了参考基因组正链中的GA变换也就是反链中的CT变换即表示它属于反链即OB。如上图即2对应的情况reads上的C全部转换成T然后mapping到了基因组正链GA转换。列个表供参考XRXG说明StrandCTCTread 是 C→TOB或者OT基因组是 C→TOT或CTOTOTCTGAread 是 C→TOB或者OT基因组是 G→AOB或CTOBOBGACTread 是 G→ACTOT或者CTOB基因组是 C→TOT或CTOTCTOTGAGAread 是 G→ACTOT或者CTOB基因组是 G→AOB或CTOBCTOB注上面这个XR和XG组合与OT、OB、CTOB、CTOT关系的对应表目前所有的AI包括GPTo3、Gemini pro、deepseek都无法得出正确答案自从本文发布后有联网搜索功能的AI查到了本文章倒是可以回答正确。通过阅读bismark源代码我们可以看到这样的注释### If the sequence aligns best as CT converted reads vs. GA converted genome (OB, index 1) or GA converted reads vs. GA converted genome (CTOB, index 3)以及### If the sequence aligns best as CT converted reads vs. CT converted genome (OT, index 0) or GA converted reads vs. CT converted genome (CTOT, index 2)验证了上表的正确性。除此之外bismark_methylation_extractor生成的六个文件CHG_OT...CHH_OT...CpG_OT...CHG_OB...CHH_OB...CpG_OB...根据bismark文档Methylation extraction - Bismark的叙述CTOT和OT的甲基化信息都被集中在OT中CTOB和OB的信息都被集中在OB中。此外这六个文件内容的格式均如下特别注意第二列的-号仅代表甲基化状态为甲基化-为非甲基化。这其实和第五列的大小写是一致的大写为甲基化小写为非甲基化即信息有冗余。全是甲基化C或者没有C的reads会直接被放进OB或者OT即使它可能来自于CTOB或CTOT### If a sequence aligns to exactly the same location with a perfect match twice the sequence does either not contain any C or G, or all the Cs (or Gs on the reverse ### strand) were methylated and therefore protected. Alternatively it will align better in one condition than in the other. In any case, it is not needed to overwrite ### the same positional entry with a second entry for the same location, as the genomic sequence extraction and methylation call would not be affected by this. The only ### thing which would change is the index number for the found alignment). We will continue to assign these alignments to the first indexes 0 and 1, i.e. OT and OB
