epigenome-profiling

What it does

Builds comprehensive epigenomic profiles for tissues or cell types using ENCODE data, guiding the user to assemble histone modification panels, assess chromatin states, run ChromHMM segmentation, identify regulatory elements, and integrate multi-omic data (histone marks, accessibility, TF binding, transcription, methylation, and 3D genome structure).

How it works

1. Choose the target biosample with encode_get_facets and assess ENCODE data availability for the tissue or cell type.
2. Assemble a histone modification panel in three tiers: Tier 1 core 5-mark panel (H3K4me3, H3K4me1, H3K27me3, H3K36me3, H3K9me3); Tier 2 extended marks (e.g., H3K9ac, H3K79me2, H2A.Z, H4K20me1, H3K27ac); Tier 3 advanced marks (e.g., H3K14ac, H3K18ac, H3K23ac, H4K5ac, H4K8ac, H4K16ac).
3. Add chromatin accessibility data prioritizing ATAC-seq; fallback to DNase-seq if needed, ensuring not to mix assays in one analysis.
4. Add transcription factor binding data, notably CTCF and EP300, via TF ChIP-seq searches; use encode_get_facets to determine available TFs for the biosample.
5. Add transcription data via polyA plus RNA-seq and/or total RNA-seq to link chromatin states to expression.
6. Add DNA methylation data with WGBS (or RRBS if WGBS unavailable), noting limitations of RRBS.
7. Add 3D chromatin data with Hi-C and/or ChIA-PET to map regulatory interactions.
8. Run ChromHMM Chromatin State Segmentation using binarized ChIP-seq signals, starting with the 5-mark core model (15 states) or 18-state extended model when H3K27ac is available.
9. Interpret states using the provided state mappings and literature references.

When to use it

• When building a comprehensive epigenomic profile for a tissue or cell type.
• When querying chromatin states, epigenome landscapes, or histone landscapes for a biosample.
• When identifying super-enhancers, bivalent domains, or regulatory elements; assembling a histone/modification/TF panel; running ChromHMM segmentation; or understanding epigenetic regulation in a context.
• When asking what ENCODE epigenomic data exists for a given tissue.

What it can touch

The skill references and suggests using these data types and tools:

• Histone ChIP-seq (for marks like H3K4me3, H3K4me1, H3K27me3, H3K36me3, H3K9me3, plus others in extended tiers)
• ATAC-seq or DNase-seq for chromatin accessibility
• TF ChIP-seq experiments (CTCF, EP300, among others discovered via encode_get_facets)
• polyA plus RNA-seq and total RNA-seq
• WGBS (or RRBS) for DNA methylation
• Hi-C and ChIA-PET for 3D genome structure
• ChromHMM for state segmentation (BinarizeBed/BinarizeBam, LearnModel, CompareModels, Reorder)

Caveats

• The workflow relies on ENCODE data availability for the chosen biosample; some assays may be missing or limited.
• RRBS covers only CpG-dense regions and is not suitable for genome-wide analyses like PMD or large HMR identification.
• The 18-state model requires H3K27ac data; otherwise, only the 15-state core model is applicable.
• Do not mix ATAC-seq and DNase-seq data within a single analysis without careful normalization.