Find clustering seeds — seedFinding • GeDi

Determine initial seeds for the clustering from the distance score matrix.

seedFinding(distances, simThreshold, memThreshold)

Arguments

distances: A Matrix::Matrix() of (distance) scores
simThreshold: numerical, A threshold to determine which genesets are considered close (i.e. have a distance <= simThreshold) in the distances matrix.
memThreshold: numerical, A threshold used to ensure that enough members of a potential seed set are close/similar to each other. Only if this condition is met, the set is considered a seed.

Value

A list of seeds which can be used for clustering

References

See https://david.ncifcrf.gov/helps/functional_classification.html#clustering for details on the original implementation

Examples

## Mock example showing how the data should look like

m <- Matrix::Matrix(stats::runif(100, min = 0, max = 1), 10, 10)
seeds <- seedFinding(distances = m, simThreshold = 0.3, memThreshold = 0.5)

## Example using the data available in the package
data(scores_macrophage_topGO_example_small,
     package = "GeDi",
     envir = environment())

seeds <- seedFinding(scores_macrophage_topGO_example_small,
                     simThreshold = 0.3,
                     memThreshold = 0.5)