Here we compute a measure of how well mixed a composite dataset is. To compute, we first examine the local neighborhood for each cell (looking at max.k neighbors) and determine for each group (could be the dataset after integration) the k nearest neighbor and what rank that neighbor was in the overall neighborhood. We then take the median across all groups as the mixing metric per cell.

```
MixingMetric(
object,
grouping.var,
reduction = "pca",
dims = 1:2,
k = 5,
max.k = 300,
eps = 0,
verbose = TRUE
)
```

- object
Seurat object

- grouping.var
Grouping variable for dataset

- reduction
Which dimensionally reduced space to use

- dims
Dimensions to use

- k
Neighbor number to examine per group

- max.k
Maximum size of local neighborhood to compute

- eps
Error bound on the neighbor finding algorithm (from RANN)

- verbose
Displays progress bar

Returns a vector of values of the mixing metric for each cell