METHimpute: imputation-guided construction of complete methylomes from WGBS data

Area of Science:

• Genomics
• Epigenetics
• Computational Biology

Background:

• Whole-genome bisulfite sequencing (WGBS) is standard for plant methylome analysis at base resolution.
• Deep WGBS is cost-prohibitive for large genomes and population studies, leading to incomplete methylome data.
• Most plant methylomes have missing data or insufficient coverage for cytosines.

Purpose of the Study:

• To develop a cost-effective computational method for complete plant methylome reconstruction.
• To enable accurate methylation analysis from low-coverage WGBS data.
• To overcome limitations of deep sequencing for large-scale genomic studies.

Main Methods:

• Developed METHimpute, a Hidden Markov Model (HMM) based imputation algorithm.
• Applied METHimpute to analyze WGBS data from maize, rice, and Arabidopsis.
• Inferred methylation status and levels for all cytosines, irrespective of sequencing coverage.

Main Results:

• METHimpute accurately infers cytosine-resolution methylomes from data as low as 6X coverage.
• The algorithm achieves high accuracy compared to 60X coverage data.
• Demonstrated cost-effectiveness for large-scale plant methylome studies.

Conclusions:

• METHimpute provides complete methylomes by imputing methylation status and levels for all cytosines.
• The method is effective even with low-coverage WGBS datasets.
• METHimpute is applicable to plants and potentially other species, with an implementation available on Bioconductor.