Ultra-Fast Analysis of Sparse DNA Methylome via Recurrent Pattern Encoding

Documentation for package ‘MethScope’ version 1.0.1

confidence_score	Produce confidence score for XGBoost prediction
confidence_score_top95	Produce confidence score based on top 95 percent for XGBoost prediction
filter_cell	Filter final prediction to reduce noise
GenerateInput	Generate pattern level data for cell type annotation
GenerateReference	Generate reference pattern labels (no default writing)
imputeRowMean	Impute missing value for 100K window matrix
Input_training	Train XGBoost model to predict cell type
nnls_deconv	Estimate cell type relative proportion
PlotConfusion	Generate confusion table for the final prediction
PlotF1	Generate F1 score barplot for each class
PlotUMAP	Generate UMAP for the final prediction based on cell patterns
PlotUMAP_fixedwindow	Generate UMAP for the final prediction based on fixed window eg.100kb bin widows
PredictCellType	Predict cell type annotation from the trained model
smooth_matrix	Smooth cell by pattern matrix to reduce noise