3. CpGtools Release History

3.1. Version 3.0.0

Major release introducing a comprehensive DNA methylation missing-value imputation framework.

New features

Added beta_impute.py, a unified command-line framework for DNA methylation missing-value analysis.
Added multiple imputation algorithms, including:
- Constant replacement
- Mean, median, minimum, and maximum imputation
- Random-value imputation
- Moving-window imputation
- K-nearest neighbors (KNN)
- Reference-based KNN
- Iterative Buck regression
- Iterative Random Forest
- SoftImpute matrix completion
- MOREL block-wise imputation
- Genomic nearest-neighbor (GNN) imputation
Added utilities for
- generating synthetic methylation matrices,
- inserting missing values into existing matrices,
- summarizing missing values, and
- evaluating imputation accuracy against a truth matrix using MAE, RMSE, and R².

Documentation

Added comprehensive documentation for beta_impute.py, including installation instructions, algorithm selection guidance, and usage examples.

3.2. Version 2.0.4

New features

Added beta_combat.py.

3.3. Version 2.0.3

Bug fixes

Fixed an issue in the ANOVA workflow where p-values and adjusted p-values were reported as missing for all CpGs.

3.4. Version 2.0.1

New features

Added predict_sex.py.
Added beta_selectNBest.py.

3.5. Version 1.10.0

New features

Added beta_UMAP.py.

3.6. Version 1.0.8

Bug fixes

Fixed an issue in beta_tSNE.py and beta_PCA.py when sample identifiers were numeric.

3.7. Version 1.0.7

New features

Added CpG_density_gene_centered.py.

3.8. Version 1.0.2

New features

Added beta_tSNE.py for t-distributed stochastic neighbor embedding (t-SNE) analysis of DNA methylation samples.

3.9. Version 1.0.1

New features

Added CpG_anno_position.py for annotating CpGs using pre-built or user-supplied genomic annotation files.

3.10. Version 1.0.0

Initial public release.