solve.Ensemble.RdGiven a TReNA object with Ensemble as the solver and a list of solvers
(default = "default.solvers"), estimate coefficients for each transcription factor
as a predictor of the target gene's expression level. The final scores for the ensemble
method combine all specified solvers to create a composite score for each transcription factor.
This method should be called using the solve method on an appropriate TReNA object.
# S4 method for EnsembleSolver run(obj)
| obj | An object of class Solver with "ensemble" as the solver string |
|---|
A data frame containing the scores for all solvers and two composite scores relating the target gene to each transcription factor. The two new scores are:
concordancea composite score
pcaMaxa composite of the principal components of the individual solver scores
Other solver methods:
run,BayesSpikeSolver-method,
run,LassoPVSolver-method,
run,LassoSolver-method,
run,PearsonSolver-method,
run,RandomForestSolver-method,
run,RidgeSolver-method,
run,SpearmanSolver-method,
run,SqrtLassoSolver-method,
run,XGBoostSolver-method
if (FALSE) { # Load included Alzheimer's data, create an Ensemble object with default solvers, and solve load(system.file(package="trena", "extdata/ampAD.154genes.mef2cTFs.278samples.RData")) target.gene <- "MEF2C" tfs <- setdiff(rownames(mtx.sub), target.gene)[1:30] ensemble.solver <- EnsembleSolver(mtx.sub, target.gene, tfs) tbl <- run(ensemble.solver) # Solve the same problem, but supply extra arguments that change alpha for LASSO to 0.8 and also # Change the gene cutoff from 10 ensemble.solver <- EnsembleSolver(mtx.sub, target.gene, tfs, geneCutoff = 0.2, alpha.lasso = 0.8) tbl <- run(ensemble.solver) # Solve the original problem with default cutoff and solver parameters, but use only 4 solvers ensemble.solver <- EnsembleSolver(mtx.sub, target.gene, tfs, solverNames = c("lasso", "pearson", "ridge")) tbl <- run(ensemble.solver) }