REGS: REGS for prediction of chemoresistance

In oncoclass/hemaClass: One-by-one normalization and classification of hematological malignacies

Description

Resistance Gene Signatures (REGS) for prediction of resistance to various chemotherapeutic drugs.

Usage

ResistanceClassifier(new.data, drugs = c("Cyclophosphamide", "Doxorubicin",
  "Vincristine", "Combined"), cut = list(Cyclophosphamide = c(0.33, 0.545),
  Doxorubicin = c(0.14, 0.9), Vincristine = c(0.46, 0.62), Combined = c(0.093,
  0.933)))

ResistancePredictor(new.data, drugs = c("Cyclophosphamide", "Doxorubicin",
  "Vincristine", "Combined"), cut = list(Cyclophosphamide = c(280, 340),
  Doxorubicin = c(280, 320), Vincristine = c(115, 127), Combined = c(200, 295)))

CyclophosphamideClassifier(new.data)

CyclophosphamidePredictor(new.data)

DoxorubicinClassifier(new.data)

DoxorubicinPredictor(new.data)

VincristineClassifier(new.data)

VincristinePredictor(new.data)

RituximabClassifier(new.data, type = "corrected", cut = c(0.33, 0.66),
  calc.cut = NULL, cut.spec = NULL, percent.classified = 85)

RituximabPredictor(new.data, type = "corrected", cut = c(0.33, 0.66),
  calc.cut = NULL)

DexamethasoneClassifier(new.data)

DexamethasonePredictor(new.data)

MelphalanClassifier(new.data)

MelphalanPredictor(new.data)

Arguments

new.data	An expression matrix.
drugs	An RMA reference object created by rmaPreprocessing.
cut	Should the .cel files be tested. When set to TRUE bad .cel files are automatically discarded.
type	For Rituximab, What type of classifier or predictor should be used. Current choices are corrected, uncorrected, lysis, and lysis2.
calc.cut	For Rituximab, calculate the cutpoints according to proportions in the data. E.g. calc.cut = c(0.33, 0.66) means that a third is deemed sensitive, intermediate, and resistant respectively.
cut.spec	For the lysis type rituximab classifier specify the cut point for unclassified.
percent.classified	For the lysis type rituximab classifier specify the percentage of unclassified.

Value

The ResistanceClassifier ResistancePredictor, RituximabClassifier, and RituximabPredictor functions return a list of length 3 with the slots:

class	A data.frame giving predicted classes for each sample.
prob	A matrix of proability or scores for each class and sample.
cut	A numeric or list of numerics giving the thresholds used to select the classes.

The remaning xxxClassifier, xxxPredictor, and xxxProbFun functions returns the components above.

Author(s)

Steffen Falgreen <sfl (at) rn.dk>
Anders Ellern Bilgrau <abilgrau (at) math.aau.dk>

References

http://hemaClass.org

See Also

rmaPreprocessing, rmaReference, microarrayScale

Examples

# First, we read the .CEL files bundled together with hemaClass
files <- list.files(system.file("extdata/celfiles", package = "hemaClass"),
                    full.names = TRUE)
affyBatch <- readCelfiles(filenames = files)

# The .CEL files are then pre-processed
affyRMA <- rmaPreprocessing(affyBatch)
# The slot exprs.sc contains median centered and scaled expression values.
# The slot exprs.sc.mean contains mean centered and scaled expression values.
# This scaling can also be achieved using the function microarrayScale.
affyRMA.sc <- microarrayScale(affyRMA$exprs, center = "median")

# We can now use the predictors

# The classifier for Cyclophosphamide, Doxorubicin, and Vincristine:
ResistanceClassifier(affyRMA.sc)

# The predictor for Cyclophosphamide, Doxorubicin, and Vincristine:
ResistancePredictor(affyRMA.sc)

# The classifier for Rituximab into Lysis, Statisk, or Resistant:
RituximabClassifier(affyRMA.sc, type = "lysis2", percent.classified = 100)

# The classifier for Rituximab into Sensitive, Intermediate, or Resistant
# without  taking human serum into account:
RituximabClassifier(affyRMA.sc, type = "uncorrected",
                    calc.cut = c(0.33, 0.66))

# The classifier for Rituximab into Sensitive, Intermediate, or Resistant
# while taking human serum into account:
RituximabClassifier(affyRMA.sc, type = "corrected")

# For the melphalan classifier we use mean centered and sd scaled expression
# values:
affyRMA.sc.mean <- microarrayScale(affyRMA$exprs, center = "mean")
MelphalanClassifier(affyRMA.sc.mean)

# For the melphalan predictor we use the original scale:
MelphalanPredictor(affyRMA$exprs)

oncoclass/hemaClass documentation built on May 24, 2019, 2:19 p.m.