The proportion of ambiguous clustering (PAC score)

The proportion of ambiguous clustering (PAC score)

PAC(consensus_mat, x1 = 0.1, x2 = 0.9, class = NULL)

Arguments

consensus_mat

A consensus matrix.

x1

Lower bound to define "ambiguous clustering".

x2

Upper bound to define "ambihuous clustering".

class

Subgroup labels. If it is provided, samples with silhouette score less than the 5^th percential are removed from PAC calculation.

Details

The PAC score is defined as F(x2) - F(x1) where F(x) is the CDF of the consensus matrix.

Value

A single numeric vaule.

See

See https://www.nature.com/articles/srep06207 for explanation of PAC score.

Author

Zuguang Gu <z.gu@dkfz.de>

Examples

data(golub_cola)
PAC(get_consensus(golub_cola[1, 1], k = 2))
#> [1] 0.2758216
PAC(get_consensus(golub_cola[1, 1], k = 3))
#> [1] 0.5309077
PAC(get_consensus(golub_cola[1, 1], k = 4))
#> [1] 0.5168232
PAC(get_consensus(golub_cola[1, 1], k = 5))
#> [1] 0.4565728
PAC(get_consensus(golub_cola[1, 1], k = 6))
#> [1] 0.3845853

# with specifying `class`
PAC(get_consensus(golub_cola[1, 1], k = 2), 
    class = get_classes(golub_cola[1, 1], k = 2)[, 1])
#> [1] 0.1874451