Compartment shift Analysis

Overview

This skill performs A/B compartment shift analysis using PC1 eigenvector values extracted from Hi-C data, following the HOMER framework. It supports two conditions, each with two or more replicates, and uses the PC1 values (E1 column) from user-provided TSV files.

Major steps include:

• Refer to Inputs & Outputs to verify necessary files.
• Always prompt user for genome assembly used. Never decide by yourself.
• Convert TSV (Chrom, start, end, weight, E1) into HOMER-compatible PC1 bedGraph files.
• Generate a unified genomic bin list for annotatePeaks.
• Extract PC1 values across all samples.
• Perform differential PC1 analysis with replicate-aware limma statistics.
• Produce differential compartment tables and stitched compartment-shift domains.

When to use this skill

Use this skill when you want to:

• Detect compartment shifts between two conditions (e.g., cell type 1 vs cell type 2)
• Identify statistically significant changes in PC1 values across genomic bins
• Determine regions that flip between A and B compartments
• Integrate compartment shift results with other genomic datasets

Inputs & Outputs

Inputs

Example input set:

• CT1_rep1.tsv
• CT1_rep2.tsv
• CT2_rep1.tsv
• CT2_rep2.tsv

Additional requirements:

• All TSVs must share identical bins.

Outputs

compartments_shift_analysis/
    shift_regions/
        diff_PC1_CT2_vs_CT1.txt
        regions.*.txt # other region files output by the tools used.
    temp/
        bins_PC1.txt
        PC1_all_samples.txt
        *.bedGraph # other bedGraph file

Decision Tree

Step 1: Convert TSV files to PC1 bedGraph

awk 'BEGIN{OFS="	"} NR>1 && NF==5 {print $1, $2, $3, $5}' CT1_rep1.tsv > CT1_rep1.PC1.bedGraph

Step 2: Create a bin list for annotatePeaks

Use any one TSV as the template:

awk 'BEGIN{OFS="	"} NR>1 && NF==5 {print $1, $2, $3}' CT1_rep1.tsv > bins_PC1.txt

The resulting bins_PC1.txt defines genomic intervals for PC1 extraction.

Step 3: Compartment shift analysis

Call:

• mcp_homer-tools__homerdifferentialPC1

with:

• binspc1path: Path to the bins_PC1.txt file generated earlier,
• genome: HOMER genome identifier, provided by user.
• bedgraphpaths: List of PC1 bedGraph files in the exact replicate order (e.g., CT1rep1, CT1rep2, CT2rep1, CT2_rep2).
• experiment_labels: List of experiment group labels matching bedGraph order (e.g. ['CT1','CT1','CT2','CT2']).
• mergedoutputpath: Output path for merged PC1 table. Empty → '<binspc1path>.merged_PC1.txt'.
• diffoutputpath: Output path for differential PC1 table. Empty → 'diff_PC1.txt'.