GeneratePredictions: Generates network scores for all genes in a network
In lancour/ignition: INTEGRATION OF GENETIC ASSOCIATIONS WITH NETWORK DIFFUSION

Description Usage Arguments Value Examples

This method takes a gene to gene distance matrix and computes the summed proximity of each gene to a set of genes. The set of genes is presumably a collection of known genes for a phenotype of interest.

1
2
3

GeneratePredictions(kernel, pheno.gene.set, autocaps = TRUE,
  invert.distance = FALSE, ignore.neighbors = FALSE, adj.matrix = NULL,
  weight.vector = NULL)

`kernel`	Required. Produced by `CreateKernel` or user may provide own gene to gene distance kernel. Row and column names must be set to gene symbols.
`pheno.gene.set`	Required. A vector of previously known phenotype-related genes. WIll automatically be coerced to uppercase unless overriden (see autocaps param).
`autocaps`	Optional. Defaults to TRUE. Converts all entries in pheno.gene.set to uppercase to ensure accurate mapping to genes in kernel.
`invert.distance`	Optional. Defaults to FALSE. If using a different kernel other than the difussion kernel produced by `CreateKernel`, this may need to be toggled to TRUE. For example, high values in a diffusion kernel are good, whereas for other distance measures, such as shortest path, low values are good.
`ignore.neighbors`	Optional. Defaults to FALSE. If set to TRUE, then predictions will only be generated for genes that do not directly interact with the pheno.gene.set. This allows for investigation of indirect interactions that can be detected with ignition's diffusion method. Note that if this is set to TRUE, an adjacency matrix such as that produced by `CreateAdjMatrix` must be provided
`adj.matrix`	Optional. An adjacency matrix of the same dimensions as the kernel. If using a custom kernel, make sure that the ith row/col of the kernel correspond to the ith row/col of the adjacency matrix. If the user has specified ignore.neighbors = TRUE, then this argument is required.
`weight.vector`	Optional. User may provide a vector of confidence measures for each gene in the specified seed set. We recommend a scale from 0 to 1, with 1 being highest possible confidence that the gene is associated with the phenotype being studied. weight.vector should be same length as pheno.gene.set

A data frame containing the computed percentiles and Z scores for each gene.

data(ignition.example.edges)
adj.matrix = CreateAdjMatrix(ignition.example.edges)
kernel = CreateKernel(adj.matrix)
known.gene.set = c('B', 'I')
GeneratePredictions(kernel, known.gene.set)