- DNA methylation sequencing and array analysis are important bioinformatics approaches used to investigate DNA methylation patterns across the genome. DNA methylation is an epigenetic modification in which a methyl group is added to DNA, most commonly to the fifth carbon of cytosine, forming 5-methylcytosine (5mC). In mammals, this modification occurs predominantly at cytosines followed by guanine nucleotides, known as CpG sites. DNA methylation can influence gene expression, chromatin organization, genomic stability, genomic imprinting, X-chromosome inactivation, and cellular differentiation. Unlike DNA sequence mutations, DNA methylation changes do not alter the underlying nucleotide sequence, but they can influence how genetic information is regulated. DNA methylation sequencing and array technologies allow researchers to measure these modifications across large portions of the genome and identify epigenetic differences between tissues, individuals, developmental stages, environmental conditions, and disease states.
- DNA methylation analysis can be performed using sequencing-based technologies or microarray platforms. Sequencing methods provide information about methylation at individual genomic positions and can cover large portions of the genome, depending on the experimental design. Array-based methods measure methylation at predefined CpG sites selected during array development. Both approaches generate quantitative methylation data, but they differ in genomic coverage, resolution, cost, sample requirements, and analytical complexity. The choice of method depends on the research objective, available biological material, genome characteristics, and the number of samples being investigated.
- Bisulfite sequencing is one of the most widely used approaches for DNA methylation analysis. In this method, genomic DNA is treated with sodium bisulfite, which converts unmethylated cytosines into uracils, while methylated cytosines are generally protected from conversion. During PCR amplification, uracils are read as thymines. After sequencing, converted and unconverted cytosines can be distinguished by comparing the sequencing reads with the reference genome. The proportion of unconverted cytosines at a particular CpG site is used to estimate its methylation level. For example, if 80 out of 100 informative reads retain cytosine at a CpG position, the estimated methylation level is approximately 80%. However, standard bisulfite sequencing does not distinguish 5-methylcytosine from 5-hydroxymethylcytosine (5hmC), another cytosine modification involved in epigenetic regulation.
- Whole-genome bisulfite sequencing (WGBS) provides comprehensive methylation information across nearly the entire genome. It involves bisulfite conversion of genomic DNA, preparation of sequencing libraries, high-throughput sequencing, and computational identification of methylated cytosines. WGBS can reveal methylation patterns in promoters, enhancers, gene bodies, intergenic regions, repetitive sequences, and other genomic elements. It is particularly useful for identifying previously unknown methylation regions and investigating genome-wide epigenetic variation. However, WGBS can be expensive because it requires substantial sequencing coverage, and bisulfite treatment may cause DNA degradation and introduce sequence-composition biases.
- Reduced representation bisulfite sequencing (RRBS) is a more targeted and cost-effective alternative to WGBS. It uses restriction enzymes, commonly MspI, to enrich DNA fragments from CpG-rich regions, including many CpG islands and gene-associated regulatory regions. These fragments undergo bisulfite conversion and sequencing. RRBS generally provides high coverage of selected CpG-rich genomic regions while reducing sequencing requirements. However, it does not provide uniform coverage of the entire genome and may miss important methylation changes in CpG-poor regions, distal enhancers, and other regulatory elements.
- Other sequencing approaches include targeted bisulfite sequencing, in which selected genomic regions are amplified and analysed, and enzymatic methyl-sequencing (EM-seq), which uses enzymatic reactions rather than harsh chemical bisulfite treatment to distinguish modified and unmodified cytosines. EM-seq can reduce DNA degradation and improve coverage in some applications. Oxidative bisulfite sequencing (oxBS-seq) and TET-assisted bisulfite sequencing (TAB-seq) provide additional information by distinguishing 5mC and 5hmC. These methods are important when researchers need to understand different forms of cytosine modification rather than measuring combined bisulfite-resistant cytosine signals.
- Long-read sequencing technologies, including Oxford Nanopore and certain PacBio platforms, provide alternative approaches for investigating DNA methylation. These technologies can detect characteristic signals associated with modified DNA bases during sequencing, potentially allowing methylation analysis without conventional bisulfite conversion. Long-read approaches can also preserve information about long-range genomic context, repetitive regions, structural variation, and allele-specific methylation. However, their accuracy and sensitivity depend on the sequencing platform, modification type, sequencing depth, and computational models used for methylation detection. They can complement short-read sequencing by providing methylation information across genomic regions that are difficult to analyse using conventional methods.
- DNA methylation arrays are widely used for large-scale epigenetic studies, particularly when many biological samples must be analysed. These platforms use probes designed to measure methylation at selected CpG sites. One of the best-known examples is the Illumina Infinium Methylation BeadChip platform, including the HumanMethylation450K and MethylationEPIC arrays. These arrays measure hundreds of thousands of CpG sites across promoters, gene bodies, CpG islands, enhancers, and other genomic regions. Each probe is designed to distinguish methylated and unmethylated DNA signals at a specific CpG site. The resulting fluorescence intensities are used to calculate methylation values. Although arrays do not cover every CpG site in the genome, they offer standardized measurements, relatively high throughput, and lower costs per sample compared with many sequencing-based approaches.
- A common measure used in methylation array analysis is the beta value, which represents the proportion of methylated signal relative to the total methylated and unmethylated signal. Beta values generally range from 0 to 1, where values near 0 indicate low methylation and values near 1 indicate high methylation. A beta value of 0.75, for example, indicates that approximately 75% of the measured signal is associated with methylated DNA, subject to platform-specific measurement characteristics. Another measure is the M-value, calculated as the logarithm of the ratio between methylated and unmethylated signals. M-values are often preferred for statistical modelling because their distribution can be more suitable for certain statistical tests, while beta values are generally easier to interpret biologically.
- DNA methylation sequencing data require specialized bioinformatics pipelines. In bisulfite sequencing, raw sequencing reads are first examined using quality-control tools such as FastQC and MultiQC. Adapter sequences and low-quality bases are removed, and reads are aligned to a bisulfite-converted reference genome using specialized tools such as Bismark or BWA-meth. Because bisulfite treatment changes unmethylated cytosines into thymines, conventional alignment approaches are not directly suitable without appropriate conversion-aware methods. After alignment, methylation levels are calculated at individual cytosine positions by comparing methylated and unmethylated read counts. Additional quality-control steps examine sequencing depth, mapping efficiency, bisulfite conversion efficiency, duplicate reads, and coverage distribution.
- For whole-genome and reduced-representation bisulfite sequencing, the resulting methylation information can be used to identify differentially methylated positions (DMPs) and differentially methylated regions (DMRs). A DMP is an individual cytosine or CpG site showing a statistically significant difference in methylation between biological groups. A DMR is a genomic region containing multiple CpG sites with coordinated differences in methylation. DMR analysis is particularly important because biologically meaningful methylation changes often involve groups of nearby CpG sites rather than isolated positions. Statistical methods must account for biological replicates, sequencing depth, variation between individuals, and potential confounding factors such as age, sex, tissue composition, and experimental batch.
- DNA methylation array analysis also requires several bioinformatics procedures. Raw intensity files are processed using software packages such as minfi, sesame, and ChAMP. These tools support sample quality assessment, background correction, normalization, probe filtering, methylation value calculation, and differential methylation analysis. Probes with poor detection quality, low signal, or known cross-reactivity may need to be excluded. Probes affected by genetic variants at the CpG site or probe-binding sequence also require careful consideration because such variants can influence measured signals independently of true methylation differences. Normalization methods help reduce technical variation between samples while preserving biologically meaningful differences.
- After preprocessing, methylation arrays can be analysed using statistical models to identify differentially methylated positions and regions. Linear models and empirical Bayes methods, including those implemented in limma, are commonly used for group comparisons. Multiple-testing correction, often using the false discovery rate (FDR), is necessary because hundreds of thousands of CpG sites may be tested simultaneously. Principal component analysis (PCA), hierarchical clustering, and heatmaps are useful for exploring sample relationships, identifying outliers, and examining patterns of methylation variation. These analyses can reveal whether samples cluster according to disease status, tissue type, developmental stage, environmental exposure, or other biological characteristics.
- An important step in DNA methylation analysis is the annotation of methylated sites and regions to genomic features. Methylation changes can occur in promoters, CpG islands, CpG island shores and shelves, gene bodies, enhancers, and intergenic regions. Promoter methylation is often associated with transcriptional repression, particularly when it affects regulatory CpG-rich regions, although the relationship depends on genomic context. Gene-body methylation can have more complex associations with transcription, alternative splicing, and transcriptional regulation. Enhancer methylation can influence the activity of distal regulatory elements. Tools and genomic annotation resources help connect methylation changes with nearby genes, regulatory regions, and biological pathways.
- Integrating DNA methylation data with other omics datasets provides a more complete understanding of gene regulation. Combining methylation profiles with RNA sequencing can reveal relationships between DNA methylation and gene expression. Integration with ATAC-seq and DNase-seq helps determine whether methylation changes are associated with chromatin accessibility. Histone modification data from ChIP-seq or CUT&Tag can provide information about active and repressive chromatin states. Genomic variant data can also be integrated to investigate methylation quantitative trait loci (meQTLs), which are genetic variants associated with differences in DNA methylation. These analyses help distinguish genetic influences on epigenetic variation from environmental and developmental effects.
- DNA methylation analysis has important applications in cancer research, where abnormal methylation patterns can contribute to gene silencing, genomic instability, altered cellular differentiation, and tumour development. Hypermethylation of certain tumour-suppressor gene promoters and global hypomethylation in repetitive genomic regions are examples of changes frequently investigated in cancer epigenetics. Methylation profiles are also studied in neurological disorders, metabolic diseases, cardiovascular conditions, immune disorders, and developmental abnormalities. In clinical research, DNA methylation signatures may support disease classification, biomarker discovery, prognosis, and treatment-response studies. However, potential biomarkers require independent validation and careful evaluation before clinical application.
- In animal genomics and breeding, DNA methylation sequencing and array analysis can help investigate the epigenetic mechanisms underlying economically important traits. Methylation differences may be associated with growth, feed efficiency, milk production, meat quality, fertility, immune function, disease resistance, and environmental adaptation. For example, methylation profiles in muscle tissue may help identify regulatory regions involved in muscle development, while mammary gland methylation patterns may provide information about milk production and mammary function. Comparisons between breeds, tissues, developmental stages, and environmental conditions can reveal epigenetic variation associated with phenotypic differences. Such information can complement quantitative genetics, genomic selection, transcriptomics, and other molecular breeding approaches.
- Environmental conditions, nutrition, stress, temperature, and management practices can influence DNA methylation patterns. These effects are particularly important in animal production because developmental and environmental exposures may affect gene regulation and phenotype. Researchers can use methylation profiling to investigate how environmental factors influence biological pathways and whether particular epigenetic patterns are associated with adaptation or production performance. However, an observed methylation difference does not automatically demonstrate that an environmental exposure caused the change or that it will be inherited by offspring. Establishing causal relationships and transgenerational epigenetic inheritance requires carefully designed experiments and appropriate evidence.
- Despite their importance, DNA methylation sequencing and array methods have several limitations. Sequencing-based approaches can be expensive, computationally demanding, and sensitive to DNA quality, sequencing depth, and conversion efficiency. Arrays are limited to predefined CpG sites and may not capture important methylation changes outside their probe coverage. Differences in cell-type composition can also influence measured methylation profiles, particularly in mixed tissues such as blood. Genetic variation, batch effects, sample handling, and normalization procedures may introduce additional biases. Therefore, robust experimental design, biological replication, appropriate statistical analysis, and independent validation are essential for reliable interpretation.
- Overall, DNA methylation sequencing and array analysis are fundamental components of modern epigenomics and bioinformatics. Sequencing approaches provide detailed information about DNA methylation across selected or extensive genomic regions, while microarrays offer standardized and cost-effective profiling of predefined CpG sites. Their integration with gene expression, chromatin accessibility, histone modifications, and genomic variation helps researchers understand how epigenetic mechanisms influence cellular function, development, disease, and complex traits. In the Bioinformatics field, these methods are essential for identifying epigenetic biomarkers, interpreting regulatory variation, and investigating the interaction between genetic information and environmental influences.