Training Data from the Top 101 Cyclic Genes in 888 Single-cell Samples

Pre-computed peco results obtained by applying fit_cyclic_many to 888 single-cell samples that have both normalized gene expression values and cell cycle labels. These training results can be used to predicting cell cycle phase in other data.

Format

A list with the following elements:

predict.yy

Estimated cyclic expression values in the training data

cellcycle_peco_ordered

Training labels ordered from 0 to 2*pi.

cell_cycle_function

Nonparametric function of cyclic gene expression trend obtained by the trendfilter function in the genlasso package.

pve

Proportion of variance explained in each gene by the cell cycle phase label.

Examples

data(training_human)