Comparative validation of oxidative bisulfite sequencing (oxBS) and chemical-assisted pyridine borane sequencing...

DNA methylation is a key regulator of tissue-specific gene expression, cell differentiation, and development. In mammals, DNA methylation predominantly occurs as 5-methylcytosine (5mC) at CpG dinucleotides. DNA methylation is a dynamic and reversible process. Increasing evidence indicates that 5-hydroxymethylcytosine (5hmC), an oxidation product of 5mC, plays important roles in chromatin organization and gene regulation. Aberrant 5hmC levels have been associated with various neurological disorders and cancer types. Despite its biological relevance, accurate quantification of 5hmC at single-CpG resolution remains challenging due to its low abundance and methodological limitations of existing detection approaches.

Here, we systematically compared bisulfite-based, bisulfite-free, and hybrid DNA conversion strategies for 5hmC analysis at single-CpG resolution with respect to accuracy, repeatability, lowest detectable (LDL) and lowest quantifiable (LQL) 5hmC levels, and DNA recovery. Method performance was evaluated using a 97 nt fragment of an MGMT enhancer containing two CpG sites. Synthetic oligonucleotides carrying cytosine, 5mC, or 5hmC at variable positions revealed that chemical-assisted pyridine borane sequencing (CAPS), a bisulfite-free approach, yielded the most accurate 5hmC quantification. However, false-positive signals resulted in increased LDL and LQL values, limiting its applicability to samples with low 5hmC abundance. Bisulfite-based protocols generally showed higher sensitivity, with the LDL and LQL values depending on the total incubation time. An optimized oxidative bisulfite protocol with a total incubation time of 140 min achieved low LDL and LQL values while substantially reducing the incubation time compared to previously published conditions. Among the oxidants tested, K₂RuO₄ resulted in more repeatable results than KRuO₄. Bisulfite-based protocols showed superior DNA recovery compared to bisulfite-free ones.

Our results demonstrate that bisulfite-based and bisulfite-free approaches for 5hmC analysis at single-CpG resolution exhibit distinct analytical strengths and limitations. Method selection should therefore be based on the specific biological question, the expected abundance of 5hmC, and practical constraints, such as DNA input and sample integrity, rather than by accuracy or sensitivity alone. It should be noted that analytical validation was primarily performed using single-stranded oligonucleotides.