sample_clusters: simple downsampling for NMF/NTF exploratory work and big...

In trichelab/velocessor: The world's dumbest single-cell RNA velocity calculator and loader

Description

Hint: if it doesn't work, you might have too many clusters. Try different options to cluster_velo_txis until you get a tranctable representation. Feel free to plot the various representations and/or look at silhouette plots, modularity, whatever until you feel comfortable that the clusters don't suck.

Usage

sample_clusters(txis, n = 100)

Arguments

txis	a SingleCellExperiment with cluster assignments in colLabels
n	how many cells per cluster to sample at most? (500)
batch	batch term (will use $sample if none)
block	block term (will omit if none)
...	additional arguments to scuttle::downsampleBatches if using

Details

And remember: Clustering Is Difficult Only When It Does Not Matter (ref 1).

Value

   a downsampled SingleCellExperiment

References

\ Daniely A, Linial N, Saks M. (2012) \ Clustering is difficult only when it does not matter. \ https://arxiv.org/abs/1205.4891

See Also

  scuttle::downsampleBatches

trichelab/velocessor documentation built on Jan. 5, 2022, 6:27 p.m.