> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/satijalab/seurat-wrappers/llms.txt
> Use this file to discover all available pages before exploring further.

# ALRAChooseKPlot()

> Visualize singular value spectrum and p-values to choose rank k for ALRA imputation

`ALRAChooseKPlot()` produces three diagnostic plots used to select the rank `k` for ALRA imputation. It requires `RunALRA()` to have been called first with `k.only = TRUE` to compute the singular value decomposition and store it in the object's tool slot.

## Syntax

```r theme={null}
ALRAChooseKPlot(
  object,
  start = 0,
  combine = TRUE
)
```

## Parameters

<ParamField path="object" type="Seurat" required>
  A Seurat object on which `RunALRA(object, k.only = TRUE)` has been run. The function reads singular values, spacings, and p-values from `Tool(object, slot = 'RunALRA')`.
</ParamField>

<ParamField path="start" type="numeric" default="0">
  Starting index for the singular value spacing and p-value plots. Defaults to `floor(k / 2)` where `k` is the automatically selected rank. Use a smaller value to see more of the spectrum.
</ParamField>

<ParamField path="combine" type="logical" default="TRUE">
  If `TRUE`, combines all three plots into a single figure using `CombinePlots()`. If `FALSE`, returns a named list of three separate ggplot objects: `spectrum`, `spacings`, and `pvals`.
</ParamField>

## Returns

When `combine = TRUE` (default): a combined ggplot object showing all three panels side by side.

When `combine = FALSE`: a named list with three ggplot objects:

* `spectrum` — singular values s\_i vs index, with a vertical line at the selected k
* `spacings` — consecutive differences s\_i - s\_{i-1}, with a vertical line at k+1
* `pvals` — p-values for each spacing, with a vertical line at k+1

## Examples

```r theme={null}
library(SeuratWrappers)

# Step 1: Compute k only (no imputation yet)
pbmc_alra <- RunALRA(pbmc_small, k.only = TRUE)

# Step 2: Visualize the singular value spectrum
ggouts <- ALRAChooseKPlot(pbmc_alra)
# Combined plot — all three panels
ggouts

# Or access individual panels
ggouts_list <- ALRAChooseKPlot(pbmc_alra, combine = FALSE)
ggouts_list$spectrum
ggouts_list$pvals

# Or use gridExtra for layout control
do.call(gridExtra::grid.arrange, c(ALRAChooseKPlot(pbmc_alra, combine = FALSE), nrow = 1))

# Step 3: Once satisfied with k, run full ALRA imputation
pbmc_alra <- RunALRA(pbmc_alra)
```

## See Also

* [`RunALRA()`](/api/run-alra) — Run ALRA imputation
* [ALRA method guide](/methods/alra)