Estimate simulation parameters for the Kersplat simulation from a real dataset. See the individual estimation functions for more details on how this is done.
Usage
kersplatEstimate(counts, params = newKersplatParams(), verbose = TRUE)
# S3 method for class 'SingleCellExperiment'
kersplatEstimate(counts, params = newKersplatParams(), verbose = TRUE)
# S3 method for class 'matrix'
kersplatEstimate(counts, params = newKersplatParams(), verbose = TRUE)
Examples
if (requireNamespace("igraph", quietly = TRUE)) {
# Load example data
library(scuttle)
set.seed(1)
sce <- mockSCE()
params <- kersplatEstimate(sce)
params
}
#> Warning: The Kersplat simulation is still experimental and may produce unreliable results. Please try it and report any issues to https://github.com/Oshlack/splatter/issues. The development version may have improved features.
#> Estimating mean parameters...
#> Selected MGE (CvM) fit
#> Estimating expression outlier parameters...
#> Estimating BCV parameters...
#> Raw: 2.6081737071714 A: 8.12109522965305 B: 1.31905398252679 C: -8.46037793603958 Y: 1.36293828975133
#> Estimating library size parameters...
#> Selected MGE (CvM) fit
#> A Params object of class KersplatParams
#> Parameters can be (estimable) or [not estimable], 'Default' or 'NOT DEFAULT'
#> Secondary parameters are usually set during simulation
#>
#> Global:
#> (GENES) (CELLS) [SEED]
#> 2000 200 787110
#>
#> 24 additional parameters
#>
#> Mean:
#> (RATE) (SHAPE) (OUT PROB)
#> 0.00272353361858179 0.475099389599897 0
#> (Out Location) (Out Scale) (DENSITY)
#> 4 0.5 Object of class density
#> [Method] [Values]
#> fit Not set
#>
#> BCV:
#> (COMMON DISP) [DoF]
#> 1.11818825375479 60
#>
#> Network:
#> [Graph] [nRegs] [regsSet]
#> Not set 100 FALSE
#>
#> Paths:
#> [nPrograms] [Means]
#> 10 Not set
#>
#> [Design]
#> data.frame (1 x 3) with columns: Path, From, Steps
#> Path From Steps
#> 1 1 0 100
#>
#> Library size:
#> (LOCATION) (SCALE) (DENSITY)
#> 12.7854772402738 0.0327578186101294 Object of class density
#> [Method]
#> fit
#>
#> Cells:
#>
#> [Design]
#> data.frame (1 x 4) with columns: Path, Probability, Alpha, Beta
#> Path Probability Alpha Beta
#> 1 1 1 1 0
#>
#> Doublets:
#> [Prop]
#> 0
#>
#> Ambient:
#> [Scale] [Empty]
#> 0.05 0
#>