AKSmooth

AKSmooth estimates single-CpG DNA methylation levels from low-coverage bisulfite sequencing (Bi-Seq) data to reconstruct methylome profiles as an alternative to high-coverage whole-genome bisulfite sequencing (WGBS).

Key Features:

• Statistical Methodology: Uses an adjusted local kernel smoother to reconstruct single-CpG methylation estimates from Bi-Seq data and is optimized for low-coverage (~4×) datasets.
• Performance and Accuracy: Achieves high concordance with high-coverage samples (Pearson correlation coefficient of 0.90) and outperforms analogous methods on low-coverage methylation profiles.
• Computational Efficiency: Demonstrates runtime performance over four and a half times faster than the reference tool.
• Application on Clinical Samples: Validated on low-coverage (~4×) DNA methylation profiles from human colon cancer samples and matched controls.
• Implementation: Implemented in R.

Scientific Applications:

• Epigenetics and DNA methylation profiling: Enables single-CpG methylation estimation for genome-wide methylome reconstruction from low-coverage Bi-Seq data.
• Oncology: Supports methylation analysis in cancer studies, exemplified by human colon cancer and matched control comparisons.
• Developmental and environmental epigenetics: Facilitates studies of epigenetic mechanisms in developmental biology and environmental epigenetics using limited sequencing depth.

Methodology:

Applies an adjusted local kernel smoother to low-coverage bisulfite sequencing (Bi-Seq) data to reconstruct single-CpG methylation estimates.