dna-methylation-file-format-handling

DNA methylation file format handling

Summary

Configure methylKit's dbtype parameter and tabix-backed storage to load bisulfite sequencing methylation calls into memory-efficient methylRawListDB objects backed by bgzipped indexed files on disk. This skill enables scalable handling of large methylation datasets without loading all data into RAM.

When to use

You have CpG methylation call files (from Bismark or MethylDackel) and need to load them into R for differential methylation analysis, but anticipate memory constraints or want to avoid loading the entire dataset into memory. Use this skill when dataset size makes in-memory storage impractical or when you need fast random access to specific genomic regions without decompressing the entire file.

When NOT to use

• Input files are already in HDF5 or other non-tabix disk-backed format (use appropriate dbtype instead)
• Your workflow requires in-memory operations only and dataset size is manageable in RAM (use default in-memory loading without dbtype parameter)
• Methylation calls are not from bisulfite sequencing or lack standard genomic coordinate format (e.g., from base-calling algorithms with non-standard output)

Inputs

• CpG methylation call files (bedGraph, cov, or Bismark coverage format)
• Sample metadata (group/phenotype assignments for methylation samples)
• Reference genome or annotation (optional, for downstream analysis)

Outputs

• methylRawListDB object with dbpath slots referencing bgzipped tabix files
• Disk-backed methylRaw objects within the list, each pointing to a tabix-indexed file
• Validated tabix file structure with methylKit metadata headers

How to apply

Call the methRead() function with the dbtype='tabix' parameter to create a methylRawListDB object that stores methylation data in bgzipped tabix-indexed files on disk rather than in memory. The resulting methylRawListDB object will have its dbpath slot populated with file paths to the bgzipped tabix files, enabling retrieval from external files. Verify that the tabix file headers contain methylKit metadata and version information (≥ 1.13.1), and inspect the object structure to confirm the dbpath slots reference valid bgzipped tabix file paths. By default, methRead() applies a minimum coverage threshold of 10 reads per base to ensure data quality, which can be adjusted as needed. This configuration allows downstream operations like unite(), calculateDiffMeth(), and getMethylDiff() to work on the database-backed object while accessing data from disk on demand.