Methylation Pipeline - Usage Guide

Overview

This workflow processes bisulfite sequencing data from FASTQ to differentially methylated regions (DMRs), covering alignment, methylation calling, and statistical analysis.

Prerequisites

# CLI tools
conda install -c bioconda bismark bowtie2 trim-galore samtools

# R packages
BiocManager::install(c('methylKit', 'genomation'))

Quick Start

Tell your AI agent what you want to do: - "Run the methylation pipeline on my bisulfite-seq data" - "Find differentially methylated regions between treatment and control" - "Align my WGBS data with Bismark and call methylation"

Example Prompts

Starting from FASTQ

"Process my RRBS data through methylation calling"

"Align bisulfite sequencing reads and extract methylation"

"Run Bismark on my whole-genome bisulfite data"

Analysis

"Find DMRs with methylKit"

"Compare methylation between tumor and normal"

"Run per-CpG differential methylation with a beta-binomial count model (DSS) instead of methylKit"

"Annotate my DMRs with gene features"

Pipeline-level decisions

"How do I check bisulfite conversion, and why does it matter before I compute methylation levels?"

"My coverage looks doubled in mate-overlap regions - did I call methylation correctly?"

"Should I use methylKit tiles or dmrseq/DSS for region-level DMRs?"

"Is a t-test on beta values good enough, or do I need a count model?"

Input Requirements

Input	Format	Description
FASTQ files	.fastq.gz	Paired-end bisulfite-treated reads
Reference	FASTA	Genome (Bismark will prepare)

What the Workflow Does

Quality Control - Trim adapters and low-quality bases
Alignment - Map bisulfite-converted reads with Bismark
Deduplication - Remove PCR duplicates
Methylation Calling - Extract methylation status per CpG
Per-CpG Analysis - Statistical testing with methylKit (R) or scipy (Python)
DMR Detection - Find differentially methylated regions

Tips

Coverage: WGBS needs 10-30x coverage; RRBS can work with less
RRBS: skip deduplication - reads legitimately stack at MspI cut sites, so positional dedup destroys real coverage (dedup WGBS/EM-seq only)
Conversion rate: Should be >99%; check with spike-in controls
M-bias: Check for position bias and trim if needed
Replicates: Minimum 2-3 per condition for reliable DMR calling
Per-CpG test choice: Sequencing counts carry coverage (precision), so route them to a beta-binomial / overdispersion count model (DSS, or methylKit overdispersion='MN'); a bare-beta t-test discards coverage and is only a quick look. Array/continuous data uses limma on M-values. See methylation-analysis/differential-cpg-testing.

methylation-analysis/bismark-alignment - Bisulfite/EM-seq alignment, library/strand model, conversion QC
methylation-analysis/methylation-calling - Per-CpG calling from BAM (Bismark/MethylDackel), contexts, variant-aware
methylation-analysis/methylkit-analysis - methylKit object model and overdispersion gotchas
methylation-analysis/differential-cpg-testing - Per-CpG testing (count-vs-continuous fork)
methylation-analysis/dmr-detection - Selection-aware region callers (dmrseq/DSS) and PMD segmentation
methylation-analysis/array-preprocessing - Alternate entry: Infinium IDAT to beta/M matrix
methylation-analysis/cell-type-deconvolution - Cell-fraction covariates for bulk-tissue EWAS
methylation-analysis/epigenetic-clocks - DNAm age and age acceleration
methylation-analysis/ewas-design - EWAS confounding, batch, inflation, and replication