Convert a partner summary table into an adjacency matrix

Usage

partner_summary2adjacency_matrix(
  x,
  sparse = TRUE,
  inputcol = "pre_id",
  outputcol = "post_id",
  inputids = NULL,
  outputids = NULL,
  standardise_input = T
)

Arguments

x: dataframe as produced by flywire_partner_summary, neuprint_connection_table or equivalent.
sparse: Whether to return a sparse matrix (default TRUE in case you are making a big one)
inputcol, outputcol: Character vector specifying the columns containing input and output identifiers. See details section for more information.
inputids, outputids: Optional vectors of input/output ids to ensure that these are present in the output matrix. Alternatively these may contain a function that takes the dataframe x as input and returns a grouping vector. See details section for more information.
standardise_input: whether to standardise the column names/types in the input dataframe. The default should work for flywire fafbseg and neuprintr input and ensure that we identify appropriate pre_id/post_id columns.

Value

A matrix with inputs as rows and outputs (downstream neurons) as columns

Details

The inputcol and outputcol arguments can name columns containing other values besides the unique numerical identifiers for neurons. For example you can refer to a cell type column, thereby generating a grouped connectivity matrix. This is very useful for bringing together neurons with similar connectivity patterns across brain hemispheres and individuals.

Passing a function to inputids and/or outputids allows partner neurons to be grouped with maximum flexibility. The input to the function will be the dataframe x (after standardisation if this has been requested). The output must be a single vector which can be interpreted as a factor to group partner neurons.

Examples

if (FALSE) {
da2ds=neuprintr::neuprint_connection_table('DA2_lPN', details=TRUE, partners='out', threshold=5)
am=partner_summary2adjacency_matrix(da2ds, inputcol = 'bodyid', outputcol = 'partner')
library(Matrix)
image(am, ylab='DA2 PNs', xlab='outputs')

amg=partner_summary2adjacency_matrix(da2ds, inputcol = 'bodyid', outputcol = 'type')
image(amg, ylab='DA2 PNs', xlab='output cell types')
}