run_protocol: Prioritizing variants using random walk with restarts
In dbottomly/HitWalker: Network-based variant prioriziation using functional assay scores

Description Usage Arguments Value Note Author(s) References See Also Examples

The random.walk function is the workhorse function for generating association scores in this implementation. The run.protocol function wraps random.walk and specifically determines the association scores for the sample variants and ranks them for downstream analysis.

1 2	run.protocol(graph.sp.mat, seed.prots, samp.vars, param.obj) random.walk(graph.sp.mat, seed.prots, rwr.params)

`graph.sp.mat`	A `dgCMatrix` from the Matrix package. Probably other types of sparse matrices would work as well, though their use is untested.
`seed.prots`	A numeric vector of hit scores (termed seeds) corresponding to nodes. The vector needs to be named by node names in `graph.sp.mat`.
`samp.vars`	A `data.frame` of variants. A column corresponding to the node IDs is necessary.
`param.obj`	A `priorDbParams` object
`rwr.params`	A `rwrParams` object

For run.protocol a list with 4 elements:

res.dta:: A data.frame containing the prioritization results
summary.id:: The column name of the node IDs
rank.col:: The column of res.dta containing the rank
type.id:: The column of res.dta containing whether the node ID corresponds to a seed (i.e. hit) or query (i.e. variant)

For random.walk a list with 3 elements:

prox.vector:: A named ordered numeric vector containing the proximity scores for all nodes in the graph
seed.prots:: The names of the hits that were present in the graph
prot.weights:: The corresponding scores of the values in seed.prots

The RWR implementation was based off of the Matlab code of Erten et al. 2011 (http://compbio.case.edu/dada/).

Daniel Bottomly

S. Erten, G. Bebek, R. Ewing and M. Koyuturk. DADA: Degree-aware algorithms for network-based disease gene prioritization. BMC BioData Mining, 4:19, 2011.

priorDbParams, rwrParams, dgCMatrix

data(prior_obj, package="HitWalker")

use.graph <- getGraph(prior.obj)
param.obj <- getParameters(prior.obj)
hit.dta <- getHitDta(prior.obj)

hit.prot.list <- split(hit.dta, hit.dta$protein)
seed.prots <- sapply(hit.prot.list, "[[", "sum.score")

graph.sp.mat <- transformGraph(param.obj)(use.graph)
samp.vars <- getVarDta(prior.obj)

run.res <- run.protocol(graph.sp.mat, seed.prots, samp.vars, param.obj)

rwr.res <- random.walk(graph.sp.mat, seed.prots, rwrParamsObj(param.obj))