Table of contents:

• Supported upstream processes
• Main workflows
• Data files

Supported upstream processes

The following upstream processes are supported as inputs to MethBat:

• methbat pileup - generates per-base-modification pileup BED files (5mC, 5hmC, 6mA) directly from aligned HiFi BAM files; its output is the input for every other pileup-consuming sub-command

Upgrading from a pre-v1.0.0 workflow that used pb-CpG-tools as the upstream pileup source? See the v1.0.0 migration guide for the full list of CLI and output-format changes.

Main workflows

There are multiple workflows that are supported by MethBat:

• Profile-based analyses - Analyses on a fixed set of pre-defined regions (for example: CpG islands for 5mC)
  • Rare methylation analysis - Identify regions in a single dataset exhibiting a "rare" methylation patterns relative to a collection of background datasets
  • Cohort methylation analysis - Identify regions exhibiting different methylation patterns between case and control datasets
• Report analysis - Analyze pre-defined regions with known expected methylation patterns (e.g., imprinting regions) and identify anomalous methylation states; requires pre-defined regions with expected methylation categories.
• Segmentation - Segment (or divide) pileup sites for an individual dataset into regions with a shared methylation pattern; no pre-defined regions required.
• Joint segmentation - Segment by averaging methylation values across a cohort at each pileup site, then segmenting the averaged values; no pre-defined regions required.
• Signature generation - Identify regions exhibiting different methylation patterns between case and control datasets; no pre-defined regions required.
• Cell type deconvolution - Estimate cell type proportions from bulk methylation data using a reference atlas; requires a pre-defined cell type atlas.