---
title: Interpreting epigenomics outputs
description: >-
  Peak tables, differential regions, methylation calls, quality metrics, and
  statistical significance for all epigenomics tools.
tags:
  - epigenomics
  - outputs
---

Each tool returns coordinate-anchored peaks, differential regions, or per-site methylation calls with statistics and quality metrics. These give you the regulatory context around a target — active enhancers and promoters, accessibility that shifts under treatment, and methylation-silenced regions — that informs target selection and where to design CRISPR guides at open chromatin.

## Interpreting epigenomics outputs

## ATAC-seq Peaks output

`atacseq_peaks` and `chipseq_peaks` return the same result shape.

| Field | Type | Description |
|-------|------|-------------|
| `num_peaks` | number | Total peaks called |
| `frip` | number | Fraction of reads in peaks (quality metric, 0-1) |
| `total_reads` | number | Total mapped reads |
| `reads_in_peaks` | number | Reads overlapping peaks |
| `method` | string | Peak caller used |
| `genome_build` | string | Reference genome build |
| `peak_width_stats` | object | `mean`, `median`, `std` peak width in base pairs |
| `chromosome_distribution` | object | Peak count per chromosome |
| `top_peaks` | array | Top peaks by signal (subset, not the full peak set) |
| `parameters` | object | Peak calling parameters used |

### Peak fields (`top_peaks`)

| Field | Type | Description |
|-------|------|-------------|
| `chrom` | string | Chromosome name |
| `start` | number | Peak start coordinate |
| `end` | number | Peak end coordinate |
| `signal_value` | number | Signal value at the peak |
| `pvalue` | number | Peak significance |
| `qvalue` | number | FDR-adjusted significance |

The result does not include gene annotation (nearest gene, distance to TSS, feature type) — chain into a genomics tool for that.

### Quality metrics

| Metric | Good | Acceptable | Poor |
|--------|------|------------|------|
| FRiP (Fraction of Reads in Peaks) | > 0.3 | 0.1 - 0.3 | < 0.1 |
| Total peaks | 30,000 - 100,000 | 10,000 - 30,000 | < 10,000 |
| Median peak width | 200 - 500 bp | 100 - 200 or 500 - 1000 bp | < 100 or > 1000 bp |

## ChIP-seq Peaks output

`chipseq_peaks` returns the same `num_peaks` / `frip` / `top_peaks` shape documented above for ATAC-seq — there is no separate `chip_reads`, `input_reads`, `signal_to_noise`, or `peak_type` field.

### Expected peak counts by mark

| Histone Mark | Typical Peak Count | Width Range |
|--------------|--------------------|-------------|
| H3K4me3 | 15,000 - 40,000 | 200 - 2,000 bp |
| H3K27ac | 20,000 - 80,000 | 200 - 3,000 bp |
| H3K4me1 | 30,000 - 100,000 | 500 - 5,000 bp |
| H3K27me3 | 5,000 - 30,000 | 5,000 - 100,000 bp |
| H3K36me3 | 10,000 - 40,000 | 5,000 - 50,000 bp |
| CTCF | 20,000 - 60,000 | 100 - 500 bp |

## ATAC-seq Differential output

| Field | Type | Description |
|-------|------|-------------|
| `differential_peaks` | number | Number of differential peaks |
| `total_peaks_tested` | number | Total peaks tested |
| `gained` | number | Peaks with increased accessibility in treatment |
| `lost` | number | Peaks with decreased accessibility in treatment |
| `method` | string | Differential method used |
| `genome_build` | string | Reference genome build |
| `fold_change_threshold` | number | Log2 fold-change threshold used |
| `fdr_threshold` | number | FDR threshold used |
| `peaks` | array | Differential peak details |

### Differential peak fields

| Field | Type | Description |
|-------|------|-------------|
| `chrom` | string | Chromosome name |
| `start` | number | Region start coordinate |
| `end` | number | Region end coordinate |
| `log2_fold_change` | number | Log2 fold change (treatment / control) |
| `p_value` | number | Raw p-value |
| `adjusted_p_value` | number | FDR-adjusted p-value |
| `treatment_count` | number | Treatment signal |
| `control_count` | number | Control signal |
| `direction` | string | `"gained"` (more open) or `"lost"` (more closed) |

The result does not include gene annotation (nearest gene, distance to TSS) — chain into a genomics tool for that.

### Fold change interpretation

| Log2 FC | Direction | Meaning |
|---------|-----------|---------|
| > 1.0 | Positive | Substantially more accessible in treatment |
| 0.5 to 1.0 | Positive | Moderately more accessible |
| -0.5 to 0.5 | -- | No significant change |
| -1.0 to -0.5 | Negative | Moderately less accessible |
| < -1.0 | Negative | Substantially less accessible |

## ChIP-seq Differential output

`chipseq_differential` returns the same shape as `atacseq_differential`, except the peak-count fields are named `differential_sites` / `total_sites_tested`, and the direction fields are `enriched` / `depleted` instead of `gained` / `lost`. Per-region `direction` values are `"enriched"` (increased binding) and `"depleted"` (decreased binding).

## Methylation Calling output

`methylation_call` returns genome-wide and per-chromosome summary statistics — it does not return a per-site table.

| Field | Type | Description |
|-------|------|-------------|
| `total_cpgs` | number | Total CpGs analyzed |
| `total_reads_processed` | number | Total reads processed |
| `mean_coverage` | number | Mean CpG coverage |
| `median_coverage` | number | Median CpG coverage |
| `global_methylation` | number | Genome-wide average methylation level (0-1) |
| `median_methylation` | number | Median methylation level (0-1) |
| `genome_build` | string | Reference genome build |
| `method` | string | Method used |
| `chromosome_methylation` | object | Mean methylation per chromosome |
| `methylation_distribution` | array | Methylation histogram bins |

### Methylation level interpretation

| Methylation | Interpretation |
|-------------|----------------|
| 0.8 - 1.0 | Heavily methylated (typically silenced) |
| 0.2 - 0.8 | Partially methylated (heterogeneous) |
| 0.0 - 0.2 | Unmethylated (typically active) |

## DMR Detection output

| Field | Type | Description |
|-------|------|-------------|
| `num_dmrs` | number | Total DMRs detected |
| `hyper` | number | Hypermethylated DMRs (treatment > control) |
| `hypo` | number | Hypomethylated DMRs (treatment < control) |
| `method` | string | DMR detection method used |
| `genome_build` | string | Reference genome build |
| `total_cpgs_tested` | number | Total CpG sites tested |
| `total_windows_tested` | number | Total sliding windows tested |
| `window_size` | number | Sliding window size (bp) |
| `methylation_diff_threshold` | number | Methylation-difference threshold used |
| `fdr_threshold` | number | FDR threshold used |
| `dmrs` | array | DMR details |

### DMR fields

| Field | Type | Description |
|-------|------|-------------|
| `chrom` | string | Chromosome name |
| `start` | number | DMR start coordinate |
| `end` | number | DMR end coordinate |
| `n_cpgs` | number | Number of CpG sites in the DMR |
| `treatment_methylation` | number | Mean methylation in treatment |
| `control_methylation` | number | Mean methylation in control |
| `methylation_difference` | number | Methylation difference (treatment - control) |
| `p_value` | number | Raw p-value |
| `adjusted_p_value` | number | FDR-adjusted p-value |
| `direction` | string | `"hyper"` or `"hypo"` |

The result does not include gene annotation — chain into a genomics tool for that.

### DMR interpretation

| Direction | Mean Diff | Biological Implication |
|-----------|-----------|----------------------|
| Hyper | > 0.2 | Gene silencing, repression gained |
| Hypo | < -0.2 | Gene activation, repression lost |

Promoter hypermethylation is commonly associated with tumor suppressor silencing in cancer. Gene body methylation often correlates with active transcription.

## Read next