icet: Identification of Cell Types using nonparametric density...
In rongmastat/IceT0.2: Identification of Cell Types using Recovered Gene Expression

Description Usage Arguments Value Author(s)

View source: R/main.R

Nonparametric density estimation is used for cell type clustering. The intermediate parameters such as bandwith and signal-to-noise ratio are chosen by build-in algorithms but can also be user-determined.

1
2
3

icet(recover.data, ratio = 20, threshold = 0.5, fdr = 0.002,
  imputation = FALSE, pca = TRUE, max_iter = 1000, perplexity = 50,
  eps = 2, MinPts = 5)

`recover.data`	The recovered or imputated gene expression data set (can be data.frame or matrix) with columns indexing the cell id and the rows indexing the genes.
`ratio`	Numeric; the signal-to-noise ratio thresholding the noise level.
`threshold`	Numeric between 0 to 1; the threshold that eliminate all the genes below that quantile in terms of mean expression levels (default: 0.5)
`fdr`	Numeric; the false discovery rate used in bandwith selection, the smaller fdr is the less genes would be selected in the end (default: 0.002)
`imputation`	A logical indicating whether the imputated data is used (default: FALSE). If 'FALSE', the list object generated from the 'recover' function in IceT package should be used.
`pca`	logical; Whether an initial PCA step should be performed (default: TRUE)
`max_iter`	integer; Number of iterations in Rtsne function (default: 1000)
`eps`	Reachability distance in dbscan function (default: 2)
`MinPts`	Reachability minimum no. of points in dbscan function (default: 5)
`numeric;`	Perplexity parameter in Rtsne function (default: 50)