`R/DownsampleMatrix.R`

`DownsampleCells.Rd`

Estimate numbers of detected genes, significantly differentially expressed genes, and median significant effect size

```
DownsampleCells(
originalData,
useAssay = "counts",
minCountDetec = 10,
minCellsDetec = 3,
minCellnum = 10,
maxCellnum = 1000,
realLabels,
depthResolution = 10,
iterations = 10,
totalReads = 1e+06
)
```

- originalData
SCtkExperiment. The SCtkExperiment object storing all assay data from the shiny app.

- useAssay
Character. The name of the assay to be used for subsampling.

- minCountDetec
Numeric. The minimum number of reads found for a gene to be considered detected.

- minCellsDetec
Numeric. The minimum number of cells a gene must have at least 1 read in for it to be considered detected.

- minCellnum
Numeric. The minimum number of virtual cells to include in the smallest simulated dataset.

- maxCellnum
Numeric. The maximum number of virtual cells to include in the largest simulated dataset

- realLabels
Character. The name of the condition of interest. Must match a name from sample data. If only two factors present in the corresponding colData, will default to t-test. If multiple factors, will default to ANOVA.

- depthResolution
Numeric. How many different read depth should the script simulate? Will simulate a number of experimental designs ranging from 10 reads to maxReadDepth, with logarithmic spacing.

- iterations
Numeric. How many times should each experimental design be simulated?

- totalReads
Numeric. How many aligned reads to put in each simulated dataset.

A 3-dimensional array, with dimensions = c(iterations, depthResolution, 3). [,,1] contains the number of detected genes in each simulated dataset, [,,2] contains the number of significantly differentially expressed genes in each simulation, and [,,3] contains the mediansignificant effect size in each simulation. If no genes are significantly differentially expressed, the median effect size defaults to infinity.

```
data("mouseBrainSubsetSCE")
subset <- mouseBrainSubsetSCE[1:1000,]
res <- DownsampleCells(subset,
realLabels = "level1class",
iterations=2)
```