Sequencing, methylation & fragmentomics

What EM-seq measures, how the genomics pipeline works, and why cfDNA fragments reflect chromatin

One library, two related layers of information

The plasma EM-seq library can describe which cytosines were methylated and—if native cfDNA fragment geometry is preserved—how the DNA molecules were fragmented. These layers are related because DNA methylation, nucleosome packaging, chromatin accessibility, cell death, and nuclease activity all influence the molecules released into plasma (Lo, Han, Jiang, and Chiu, 2021).

From nucleosome-protected cfDNA to EM-seq reads and a joint methylation–fragmentomics profile.

What is actually sequenced?

The sequencer reads both ends of short DNA molecules in an EM-seq library. For each paired-end plasma molecule, the analysis may recover:

Observable	Derived from	Biological layer
Read sequence and genomic position	Bases and alignment to GRCh38	Genomic origin of the molecule
Cytosine state	Enzymatic conversion followed by sequencing	5mC/5hmC protection versus unmodified cytosine conversion
Fragment length	Distance and orientation of a properly paired alignment	Size distribution and nucleosomal periodicity
Fragment ends	Left/right genomic coordinates and terminal sequence motifs	Chromatin- and nuclease-associated cleavage patterns
Regional coverage	Counts of overlapping fragments	Genome-wide representation and footprints near regulatory regions

Fragmentomics does not sequence a separate molecule type. It reuses positional and physical information carried by paired-end reads from native plasma cfDNA. Tissue genomic DNA is deliberately sheared for library preparation, so its insert lengths describe the laboratory fragmentation step and should not be interpreted as native tissue fragmentomics.

Why histones matter

DNA in the nucleus is wrapped around histone proteins to form nucleosomes. During cell death and DNA release, nucleosome-protected DNA is less accessible to nucleases than linker DNA. Plasma cfDNA therefore often shows a mono-nucleosomal mode near 167 bp and longer multiples, plus coverage and end-position patterns related to chromatin organization (Lo, Han, Jiang, and Chiu, 2021).

flowchart LR
  C["Cell nucleus"] --> N["DNA wrapped around<br/>histone nucleosomes"]
  N --> D["Cell death +<br/>nuclease cleavage"]
  D --> F["Protected cfDNA<br/>fragments in plasma"]
  F --> L["Length + ends +<br/>coverage footprints"]
  F --> M["Methylation state<br/>along the fragment"]
  L --> I["Joint biological<br/>interpretation"]
  M --> I

Histones are not directly sequenced by EM-seq. Their organization is inferred indirectly from patterns in DNA fragment size, positions, and coverage. This inference can be altered by collection delay, nuclease activity, extraction, end repair, conversion, PCR, sequencing depth, and mapping.

How methylation and fragmentation connect

Methylation is associated with chromatin state, and cleavage near CpGs can differ with methylation configuration. FRAGMA demonstrated methylation-aware CpG cleavage profiles (Zhou, Kang, Jiang, and others, 2022), while FinaleMe used fragment length, coverage, and CpG position within a fragment to infer methylation from plasma WGS (Liu, Reed, Lo, Choudhury, and others, 2024). These papers justify testing linked features, but their trained models are not automatically valid for this EM-seq protocol or UC cohort.

NEBNext EM-seq wet-lab logic

NEBNext Enzymatic Methyl-seq avoids the harsh bisulfite reaction and supports low-input, fragmented DNA (Vaisvila, Ponnaluri, Sun, and others, 2021). In simplified terms:

1. Prepare the DNA library. Preserve native plasma cfDNA; shear tissue DNA in a controlled, documented step.
2. Protect modified cytosines. TET2 oxidation and T4-BGT chemistry protect 5mC and 5hmC from the later deamination step.
3. Convert unmodified cytosines. APOBEC deaminates unmodified cytosine so it is read as thymine after amplification and sequencing.
4. Sequence paired ends. Illumina reads retain methylation information and, for plasma, the paired alignment needed to estimate insert length and ends.

NoteWhat the assay reports

Standard EM-seq readout generally represents protected 5mC and 5hmC together rather than separating them. Interpret it as modified-cytosine signal unless an additional assay distinguishes the two.

Genomics pipeline

flowchart TD
  R["Paired FASTQ"] --> Q["fastp<br/>adapter + quality QC"]
  Q --> A["bwa-meth<br/>alignment to GRCh38"]
  A --> B["Coordinate-sorted BAM<br/>pair + insert QC"]
  B --> MC["MethylDackel<br/>CpG methylation calls"]
  MC --> DS["DSS<br/>DML/DMR models"]
  DS --> PA["Gene mapping +<br/>KEGG/Reactome pathways"]
  B --> FR["Fragment feature table<br/>lengths, ends, TSS coverage"]
  MC --> JM["Methylation × length<br/>joint summaries"]
  FR --> JM
  PA --> IN["Integrated model"]
  JM --> IN
  S["sIgA + inflammation<br/>+ clinical covariates"] --> IN

Protocol-by-protocol guides

Stage	Guide
18 consented participants · three groups	Cohort, consent, and study manifest
Blood → plasma + serum	Blood processing and aliquots
Colonic biopsy → prepared tissue	Colonic biopsy preparation
PANAMAX 48 · plasma	Automated cfDNA extraction
PANAMAX 48 · tissue	Automated genomic-DNA extraction
UltraShear	Tissue-DNA fragmentation
NEBNext EM-seq	Methylation library preparation
NovaSeq 6000	Paired-end sequencing
Serum sIgA	ELISA processing and data

Minimum outputs and checks

Stage	Required output	Stop/go question
FASTQ QC	Read counts, base quality, adapter content, duplication indicators	Is the library usable without differential quality by group?
Alignment	Mapping rate, proper-pair rate, duplicates, conversion controls, insert sizes	Are positions, methylation calls, and native plasma sizes credible?
Methylation	Coverage distribution, conversion performance, CpG calls	Is coverage sufficient and comparable for DML/DMR modeling?
Fragmentomics	Length distribution, short ratio, nucleosomal peaks, ends, TSS coverage	Was native fragment structure preserved and is depth adequate?
Integration	Locked feature table with batch and clinical covariates	Can uncertainty be estimated without feature leakage or overfitting?

Pathway interpretation

After FDR-controlled DML/DMR analysis, mapped genes can be summarized with KEGG and Reactome. The report should retain the CpG/DMR-to-gene rule, tested-region background, database version, effect direction, overlap size, and adjusted p-value. Enrichment is a structured hypothesis about biological context—not direct evidence that a pathway is active.

Learn what KEGG contributes → · Explore Reactome’s reaction hierarchy →

ImportantCritical boundary

With 18 planned participants, the defensible target is a quality-controlled feasibility analysis and a small set of candidate features. High-dimensional classifiers, prognosis claims, and clinical use require independent longitudinal validation.

See the complete experimental roadmap → · Read the literature review →

Prefer a visual introduction? Read “What is enzymatic methyl sequencing?” →

Then explore “What is fragmentomics?” →

See how the analysis software evolved → · Understand beta-binomial and regional statistics →

References

Y. M. Dennis Lo, Diana S. C. Han, Peiyong Jiang, and Rossa W. K. Chiu. (2021). Epigenetics, fragmentomics, and topology of cell-free DNA in liquid biopsies. Science. https://doi.org/10.1126/science.aaw3616.

Qing Zhou, Guannan Kang, Peiyong Jiang, and others. (2022). Epigenetic analysis of cell-free DNA by fragmentomic profiling. Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.2209852119.

Yaping Liu, Sarah C. Reed, Christopher Lo, Atish D. Choudhury, and others. (2024). FinaleMe: Predicting DNA methylation by the fragmentation patterns of plasma cell-free DNA. Nature Communications. https://doi.org/10.1038/s41467-024-47196-6.

Romualdas Vaisvila, Vamsi K. C. Ponnaluri, Zhe Sun, and others. (2021). Enzymatic methyl sequencing detects DNA methylation at single-base resolution from picograms of DNA. Genome Research. https://doi.org/10.1101/gr.266551.120.