coexpression_known_interactions_precision_at_top: Precision of a co-expression network.

In alorchhota/spice: SPICE: Spanning tree based inference of co-expression networks

Description

This function computes the precision of a co-expression network at top edges i.e., the fraction of true edges in a given number of top weighted edges. It has options to set the number of top edges and the known score thresholds to define which edges are true.

Usage

coexpression_known_interactions_precision_at_top(
  net,
  known,
  score.thresholds,
  n.top.edges = NA,
  na.ignore = "known",
  neg.treat = "error"
)

Arguments

net	matrix or data.frame. A gene x gene matrix representing edge weights between genes in a co-expression network. See details.
known	matrix or data.frame. A gene x gene matrix representing the probability that edges between genes true. See details.
score.thresholds	numeric vector. Each value must be in the rage [0,1]. If known interaction score is equal to or greater than a score threshold, the corresponding edge is considered true.
n.top.edges	vector with numeric values or NA. The number of top weighted edges in net to use to compute the precision. If NA, the number of edges with a known interaction score equal to or greater than the given score threshold is used. If the number of edges is greater than the number of available edges, all available edges are used. If multiple edges have the same weight as the last top-weighted edge, all of those edges are used.
na.ignore	character representing how NA's should be handled. Accepted values are 'net', 'known' and 'any'. If 'net', edges with NA weight in net are ignored. If 'known', edges with NA weight in known are ignored. If 'any', edges with NA weight in either net or known are ignored.
neg.treat	character representing how negative values in net should be treated. Accepted values are 'none', 'warn' and 'error'. If 'allow', negative values are allowed. If 'warn', a warning is generated. If 'error', an error is generated.

Details

Each value in known must be in the range [0, 1] representing the probability that the corresponding edge (interaction) is true. While the values in net are not limited to any range, each value should represent the relative probability that the corresponding edge is true. In other words, larger values should represent higher confidence in corresponding edges. If the sign of values in net represents positive or negative associations between genes, you probably should provide absolute values. If you still want to allow negative values in net, you may set neg.treat = "allow". In this case, any negative value will represent lower confidence than any non-negative value.

Both net and known must be square matrices of same dimension. Either both or none of the matrices should have row and column names. If available, the set of row names and column names must be unique and same in each matrix. The set of row and columns names of both matrices should also be same. Both matrices must be symmetric when rows and columns are in the same order. Diagonal entries in the matrices are ignored.

Value

A matrix with length(score.thresholds) rows and length(n.top.edges) columns. Each value in the matrix represent the precision at a given number of top edges (column) using a given score threshold (row).

Examples

genes = sprintf("G%d", 1:10)
dummy_net = matrix(rnorm(length(genes)^2), nrow = length(genes), dimnames = list(genes, genes))
dummy_net = abs((dummy_net + t(dummy_net))/2)                    # symmetric network
dummy_ppi = abs(dummy_net + rnorm(length(dummy_net)))
dummy_ppi = (dummy_ppi + t(dummy_ppi)) / (2 * max(dummy_ppi))    # symmetric ppi
net_precision = coexpression_known_interactions_precision_at_top(
  net = dummy_net,
  known = dummy_ppi,
  n.top.edges = c(10, 30, NA),
  score.thresholds = c(0.25, 0.5)
)
print(net_precision)

alorchhota/spice documentation built on March 12, 2021, 12:05 a.m.