connectivityCluster: Clustering neurons based on connectivity tables

In jayaraman-lab/neuprintrExtra: Extra functions to analyze neuprint EM datasets

Description

Clustering neurons based on connectivity tables

Usage

connectivityCluster(
  inputsTable = NULL,
  outputsTable = NULL,
  ROIs = NULL,
  grouping = "type",
  distance = cos_dist,
  knownStats = FALSE,
  unitNames = NULL,
  ...
)

clusterBag(
  nBag,
  ROIs = NULL,
  clusterOn = c("inputs", "outputs", "both"),
  grouping = c("type", "neuron", "bodyid", "supertype1", "supertype2", "supertype3",
    "databaseType"),
  distance = cos_dist,
  knownStats = FALSE,
  ...
)

Arguments

inputsTable	The table of input connections to the neurons to cluster
outputsTable	The table of input connections to the neurons to cluster
ROIs	Which ROIs to consider. If NULL (the default) uses all the ROI present
grouping	grouping on which the clustering is to be made (neuron or type, or various supertypes)
distance	Which distance measure to use (defaults to cos_dist)
knownStats	Whether or not to use knownOutputContribution and knownWeightRelative instead of the standard measures
unitNames	Optional vector argument of the neurons/types one wants to cluster
...	Arguments to be passed to hclust
nBag	a neuron bag
clusterOn	whether to cluster on the input vector, the output vector, or both

Details

• If both inputsTable and outputsTable are present, the clustering will be run on the combined connectivity vector. This is controlled by the clusterOn argument in clusterBag.

• weightRelative is used for working on input vectors, and outputContribution to work on output vectors

Value

an object of class connectivityCluster. The object is a list with components:

inputsTable

The table of input connections to the neurons being clustered (can be NULL if clustered on outputs)

outputsTable

The table of output connections to the neurons being clustered (can be NULL if clustered on inputs)

distance

object containing a dissimilarity measure between the neurons being considered

hc

hclust object, result of clustering

grouping

grouping on which the clustering was made (neuron or type, or various supertypes)

Functions

• clusterBag: Cluster neurons in a bag

See Also

clusterBag

jayaraman-lab/neuprintrExtra documentation built on Dec. 20, 2021, 10 p.m.