1. edge
  2. Commit 9001e378c0bb8098fa8669d0db7a051d14b93d16
Browse code

add packages to the repository

edge/ pwOmics/ EMDomics/



git-svn-id: file:///home/git/hedgehog.fhcrc.org/bioconductor/trunk/madman/Rpacks/edge@102444 bc3139a8-67e5-0310-9ffc-ced21a209358

Sonali Arora authored on 14/04/2015 17:06:48
Showing 59 changed files
DESCRIPTION
History View file @ 9001e37
1 1 
new file mode 100644
... ... 
@@ -0,0 +1,34 @@
1 
+Package: edge
2 
+Type: Package
3 
+Title: Extraction of Differential Gene Expression
4 
+Date: 2015-03-26
5 
+Version: 0.99.4
6 
+Authors@R: c( person("John D.", "Storey", email =
7 
+        "jstorey@princeton.edu", role = c("aut", "cre", "cph")),
8 
+        person("Jeffrey T.", "Leek", email = "jleek@jhsph.edu ", role =
9 
+        c("aut")), person("Andrew J.", "Bass", email =
10 
+        "ajbass@princeton.edu", role = c("aut")) )
11 
+Maintainer: John D. Storey <jstorey@princeton.edu>
12 
+biocViews: MultipleComparison, DifferentialExpression, TimeCourse,
13 
+        Regression, GeneExpression, DataImport
14 
+Description: The edge package implements methods for carrying out
15 
+        differential expression analyses of genome-wide gene expression
16 
+        studies. Significance testing using the optimal discovery
17 
+        procedure and generalized likelihood ratio tests (equivalent to
18 
+        F-tests and t-tests) are implemented for general study designs.
19 
+        Special functions are available to facilitate the analysis of
20 
+        common study designs, including time course experiments. Other
21 
+        packages such as snm, sva, and qvalue are integrated in edge to
22 
+        provide a wide range of tools for gene expression analysis.
23 
+VignetteBuilder: knitr
24 
+Imports: methods, splines, sva, snm, qvalue(>= 1.99.0), MASS
25 
+Suggests: testthat, knitr, ggplot2, reshape2
26 
+Depends: R(>= 3.2.0), Biobase
27 
+URL: https://github.com/jdstorey/edge
28 
+BugReports: https://github.com/jdstorey/edge/issues
29 
+LazyData: true
30 
+License: GPL
31 
+Copyright: 2005-2015 John D. Storey
32 
+NeedsCompilation: yes
33 
+Author: John D. Storey [aut, cre, cph], Jeffrey T. Leek [aut], Andrew
34 
+        J. Bass [aut]
NAMESPACE
History View file @ 9001e37
0 35 
new file mode 100644
... ... 
@@ -0,0 +1,43 @@
1 
+# Generated by roxygen2 (4.1.0): do not edit by hand
2 
+
3 
+export("fullMatrix<-")
4 
+export("fullModel<-")
5 
+export("individual<-")
6 
+export("nullMatrix<-")
7 
+export("nullModel<-")
8 
+export("qvalueObj<-")
9 
+export(apply_qvalue)
10 
+export(apply_snm)
11 
+export(apply_sva)
12 
+export(betaCoef)
13 
+export(build_models)
14 
+export(build_study)
15 
+export(deSet)
16 
+export(fitFull)
17 
+export(fitNull)
18 
+export(fullMatrix)
19 
+export(fullModel)
20 
+export(individual)
21 
+export(lrt)
22 
+export(nullMatrix)
23 
+export(odp)
24 
+export(qvalueObj)
25 
+export(resFull)
26 
+export(resNull)
27 
+export(show)
28 
+export(summary)
29 
+exportClasses(deFit)
30 
+exportClasses(deSet)
31 
+exportMethods(fit_models)
32 
+exportMethods(kl_clust)
33 
+exportMethods(nullModel)
34 
+exportMethods(sType)
35 
+import(Biobase)
36 
+import(MASS)
37 
+import(methods)
38 
+import(qvalue)
39 
+import(snm)
40 
+import(splines)
41 
+import(sva)
42 
+useDynLib(edge,kldistance)
43 
+useDynLib(edge,odpScoreCluster)
R/AllClasses.R
History View file @ 9001e37
0 44 
new file mode 100644
... ... 
@@ -0,0 +1,201 @@
1 
+# Allows to set qvalue to S4 slot
2 
+setOldClass("qvalue")
3 
+
4 
+deSetCheck <- function(object) {
5 
+  errors <- character()
6 
+  epsilon <- 10e-8
7 
+  # Allow easy conversion for an ExpressionSet using function 'as'
8 
+  if (is.list(object@null.model) && is.list(object@full.model) &&
9 
+        length(object@individual) == 0) {
10 
+    return(TRUE)
11 
+  }
12 
+  # Name mismatch
13 
+  f.vars <- all.vars(object@full.model)
14 
+  n.vars <- all.vars(object@null.model)
15 
+  names <- unique(c(f.vars, n.vars))
16 
+  if (sum((f.vars %in% c("grp", "bio.var", "time.basis"))) == 0) {
17 
+    if (sum(!(names %in% varLabels(object))) != 0) {
18 
+      msg <- paste("naming mismatch between phenoData covariates and models.")
19 
+      errors <- c(errors, msg)
20 
+      return(errors)
21 
+    }
22 
+  }
23 
+  # Singular matrix
24 
+  xx0 <- model.matrix(object@null.model, data=object)
25 
+  xx1 <- model.matrix(object@full.model, data=object)
26 
+  #  sCheck.null <- min(svd(xx0)$d) < epsilon
27 
+  sCheck.full <- min(svd(xx1)$d) < epsilon
28 
+  #  if (sCheck.null) {
29 
+  #   msg <- paste("null model matrix is near singular.")
30 
+  #   errors <- c(errors, msg)
31 
+  #  }
32 
+  if (sCheck.full) {
33 
+    msg <- paste("full model matrix is near singular.")
34 
+    errors <- c(errors, msg)
35 
+  }
36 
+  # Dimensionality test- this may be impossible to make in deSet
37 
+  dataDim <- dim(exprs(object))
38 
+  if (dataDim[2] != nrow(xx1)) {
39 
+    msg <- paste( "dimension mismatch between full model and assayData.")
40 
+    errors <- c(errors, msg)
41 
+  }
42 
+  if (dataDim[2] != nrow(xx0)) {
43 
+    msg <- paste( "dimension mismatch between null model and assayData.")
44 
+    errors <- c(errors, msg)
45 
+  }
46 
+  # inidividual input size
47 
+  if (length(object@individual) != 0) {
48 
+    if (length(object@individual) != ncol(exprs(object))) {
49 
+      msg <- paste("individual must be the same length as the number of arrays")
50 
+      errors <- c(errors, msg)
51 
+    }
52 
+  }
53 
+  if (length(errors) == 0) {
54 
+    TRUE
55 
+  } else {
56 
+    errors
57 
+  }
58 
+}
59 
+
60 
+deFitCheck <- function(object) {
61 
+  errors <- character()
62 
+  # Dimensionality test
63 
+  if (!(    (ncol(object@fit.full)==ncol(object@fit.null)
64 
+             && (ncol(object@res.full) == ncol(object@res.null))
65 
+             && (length(object@dH.full) == ncol(object@fit.full))
66 
+             && (ncol(object@fit.full) == ncol(object@res.null))))) {
67 
+    msg <- paste("column length of fitted matrices, dH.full and residuals",
68 
+                 "must be the same.")
69 
+    errors <- c(errors, msg)
70 
+  }
71 
+  if (!((nrow(object@fit.full) == nrow(object@fit.null))
72 
+        && (nrow(object@res.full) == nrow(object@res.null))
73 
+        && (nrow(object@res.full) == nrow(object@fit.full))
74 
+        && (nrow(object@beta.coef) == nrow(object@fit.null)))) {
75 
+    msg <- paste("row length of fitted matrices and residuals",
76 
+                 "must be the same.")
77 
+    errors <- c(errors, msg)
78 
+  }
79 
+  # Correct statistic input check
80 
+  if (!(object@stat.type %in% c("lrt", "odp"))) {
81 
+    msg <- paste("stat.type must be lrt or odp. Inputted stat.type: ",
82 
+                 object@stat.type)
83 
+    errors <- c(errors, msg)
84 
+  }
85 
+  if (length(errors) == 0) {
86 
+    TRUE
87 
+  } else {
88 
+    errors
89 
+  }
90 
+}
91 
+
92 
+#' The differential expression class (deSet)
93 
+#'
94 
+#' The \code{deSet} class is designed in order to complement the
95 
+#' \code{\link{ExpressionSet}} class. While the \code{ExpressionSet} class
96 
+#' contains information about the experiment, the \code{deSet} class
97 
+#' contains both experimental information and additional information relevant
98 
+#' for differential expression analysis.
99 
+#'
100 
+#' @slot null.model \code{formula}: contains the adjustment variables in the
101 
+#' experiment. The null model is used for comparison when fitting the
102 
+#' full model.
103 
+#' @slot full.model \code{formula}: contains the adjustment variables and the
104 
+#' biological variables of interest.
105 
+#' @slot null.matrix \code{matrix}: the null model as a matrix.
106 
+#' @slot full.matrix \code{matrix}: the full model as a matrix.
107 
+#' @slot individual \code{factor}: contains information on which sample
108 
+#' is from which individual in the experiment.
109 
+#' @slot qvalueObj \code{S3 object}: containing \code{qvalue} object.
110 
+#' See \code{\link{qvalue}} for additional details.
111 
+#'
112 
+#' @section Methods:
113 
+#'  \describe{
114 
+#'  \item{\code{as(ExpressionSet, "deSet")}}{Coerce objects of
115 
+#'  \code{ExpressionSet} to \code{deSet}.}
116 
+#'  \item{\code{lrt(deSet, ...)}}{Performs a generalized likelihood ratio test
117 
+#'  using the full and null models.}
118 
+#'  \item{\code{odp(deSet, ...)}}{Performs the optimal discovery procedure,
119 
+#'  which is a new approach for optimally performing many hypothesis tests in
120 
+#'  a high-dimensional study.}
121 
+#'  \item{\code{kl_clust(deSet, ...)}}{An implementation of mODP that assigns
122 
+#'  genes to modules based off of the Kullback-Leibler distance.}
123 
+#'  \item{\code{fit_models(deSet, ...)}}{Fits a linear model to each gene by
124 
+#'  method of least squares.}
125 
+#'  \item{\code{apply_qvalue(deSet, ...)}}{Applies \code{\link{qvalue}}
126 
+#'  function.}
127 
+#'  \item{\code{apply_snm(deSet, ...)}}{Applies surpervised normalization of
128 
+#'   microarrays (\code{\link{snm}}) on gene expression data.}
129 
+#'  \item{\code{apply_sva(deSet, ...)}}{Applies surrogate variable analysis
130 
+#'  (\code{\link{sva}}).}
131 
+#'  \item{\code{fullMatrix(deSet)}}{Access and set full matrix.}
132 
+#'  \item{\code{nullMatrix(deSet)}}{Access and set null matrix.}
133 
+#'  \item{\code{fullModel(deSet)}}{Access and set full model.}
134 
+#'  \item{\code{nullModel(deSet)}}{Access and set null model.}
135 
+#'  \item{\code{individual(deSet)}}{Access and set individual slot.}
136 
+#'  \item{\code{qvalueObj(deSet)}}{Access \code{qvalue} object.
137 
+#'  See \code{\link{qvalue}}.}
138 
+#'  \item{\code{validObject(deSet)}}{Check validity of \code{deSet} object.}
139 
+#'  }
140 
+#'
141 
+#' @note
142 
+#' See \code{\link{ExpressionSet}} for additional slot information.
143 
+#'
144 
+#' @author
145 
+#' John Storey, Jeffrey Leek, Andrew Bass
146 
+#' @inheritParams ExpressionSet
147 
+#' @exportClass deSet
148 
+setClass("deSet", slots=c(null.model = "formula",
149 
+                          full.model = "formula",
150 
+                          null.matrix = "matrix",
151 
+                          full.matrix = "matrix",
152 
+                          individual = "factor",
153 
+                          qvalueObj = "qvalue"),
154 
+         prototype=prototype(null.model = formula(NULL),
155 
+                             full.model = formula(NULL),
156 
+                             null.matrix = matrix(),
157 
+                             full.matrix = matrix(),
158 
+                             individual = as.factor(NULL),
159 
+                             qvalueObj = structure(list(),
160 
+                                                    class = "qvalue")),
161 
+         validity = deSetCheck,
162 
+         contains = c("ExpressionSet"))
163 
+
164 
+#' The differential expression class for the model fits
165 
+#'
166 
+#' Object returned from \code{\link{fit_models}} containing information
167 
+#' regarding the model fits for the experiment.
168 
+#'
169 
+#' @slot fit.full \code{matrix}: containing fitted values for the full model.
170 
+#' @slot fit.null \code{matrix}: containing fitted values for the null model.
171 
+#' @slot res.full \code{matrix}: the residuals of the full model.
172 
+#' @slot res.null \code{matrix}: the residuals of the null model.
173 
+#' @slot dH.full \code{vector}: contains diagonal elements in the projection
174 
+#' matrix for the full model.
175 
+#' @slot beta.coef \code{matrix}: fitted coefficients for the full model.
176 
+#' @slot stat.type \code{string}: information on the statistic of interest.
177 
+#' Currently, the only options are ``lrt'' and ``odp''.
178 
+#'
179 
+#' @section Methods:
180 
+#'  \describe{
181 
+#'  \item{\code{fitNull(deFit)}}{Access fitted data from null model.}
182 
+#'  \item{\code{fitFull(deFit)}}{Access fitted data from full model.}
183 
+#'  \item{\code{resNull(deFit)}}{Access residuals from null model fit.}
184 
+#'  \item{\code{resFull(deFit)}}{Access residuals from full model fit.}
185 
+#'  \item{\code{betaCoef(deFit)}}{Access beta coefficients in linear model.}
186 
+#'  \item{\code{sType(deFit)}}{Access statistic type of model fitting utilized
187 
+#'  in function.}
188 
+#'  }
189 
+#'
190 
+#' @author
191 
+#' John Storey, Jeffrey Leek, Andrew Bass
192 
+#'
193 
+#' @exportClass deFit
194 
+setClass("deFit", slots=c(fit.full = "matrix",
195 
+                          fit.null = "matrix",
196 
+                          res.full = "matrix",
197 
+                          res.null = "matrix",
198 
+                          dH.full = "vector",
199 
+                          beta.coef = "matrix",
200 
+                          stat.type = "character"),
201 
+         validity = deFitCheck)
R/AllGenerics.R
History View file @ 9001e37
0 202 
new file mode 100644
... ... 
@@ -0,0 +1,1116 @@
1 
+#' Performs F-test (likelihood ratio test using Normal likelihood)
2 
+#'
3 
+#' \code{lrt} performs a generalized likelihood ratio test using the full and
4 
+#' null models.
5 
+#'
6 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}.
7 
+#' @param de.fit \code{S4 object}: \code{\linkS4class{deFit}}. Optional.
8 
+#' @param nullDistn \code{character}: either "normal" or "bootstrap", If
9 
+#' "normal" then the p-values are calculated using the F distribution. If
10 
+#' "bootstrap" then a bootstrap algorithm is implemented to simulate
11 
+#' statistics from the null distribution. In the "bootstrap" case, empirical
12 
+#' p-values are calculated using the observed and null statistics (see
13 
+#' \code{\link{empPvals}}). Default is "normal".
14 
+#' @param bs.its \code{integer}: number of null statistics generated (only
15 
+#' applicable for "bootstrap" method). Default is 100.
16 
+#' @param seed \code{integer}: set the seed value. Default is NULL.
17 
+#' @param verbose \code{boolean}: print iterations for bootstrap method.
18 
+#' Default is TRUE.
19 
+#' @param ... Additional arguments for \code{\link{apply_qvalue}} and
20 
+#' \code{\link{empPvals}} function.
21 
+#'
22 
+#' @details
23 
+#' \code{lrt} fits the full and null models to each gene using the function
24 
+#' \code{\link{fit_models}} and then performs a likelihood ratio test. The
25 
+#' user has the option to calculate p-values from either the F distribution or
26 
+#' through a bootstrap algorithm. If \code{nullDistn} is "bootstrap"
27 
+#' then empirical p-values will be determined from the \code{\link{qvalue}}
28 
+#' package (see \code{\link{empPvals}}).
29 
+#'
30 
+#' @author John Storey, Andrew Bass
31 
+#'
32 
+#' @return \code{\linkS4class{deSet}} object
33 
+#'
34 
+#' @examples
35 
+#' # import data
36 
+#' library(splines)
37 
+#' data(kidney)
38 
+#' age <- kidney$age
39 
+#' sex <- kidney$sex
40 
+#' kidexpr <- kidney$kidexpr
41 
+#' cov <- data.frame(sex = sex, age = age)
42 
+#'
43 
+#' # create models
44 
+#' null_model <- ~sex
45 
+#' full_model <- ~sex + ns(age, df = 4)
46 
+#'
47 
+#' # create deSet object from data
48 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
49 
+#' full.model = full_model)
50 
+#'
51 
+#' # lrt method
52 
+#' de_lrt <- lrt(de_obj, nullDistn = "normal")
53 
+#'
54 
+#' # to generate p-values from bootstrap
55 
+#' de_lrt <- lrt(de_obj, nullDistn = "bootstrap", bs.its = 30)
56 
+#'
57 
+#' # input a deFit object directly
58 
+#' de_fit <- fit_models(de_obj, stat.type = "lrt")
59 
+#' de_lrt <- lrt(de_obj, de.fit = de_fit)
60 
+#'
61 
+#' # summarize object
62 
+#' summary(de_lrt)
63 
+#'
64 
+#' @references
65 
+#' Storey JD, Xiao W, Leek JT, Tompkins RG, and Davis RW. (2005) Significance
66 
+#' analysis of time course microarray experiments. Proceedings of the National
67 
+#' Academy of Sciences, 102: 12837-12842.
68 
+#'
69 
+#' @seealso \code{\linkS4class{deSet}}, \code{\link{build_models}},
70 
+#' \code{\link{odp}}
71 
+#'
72 
+#' @export
73 
+setGeneric("lrt", function(object, de.fit,
74 
+                           nullDistn = c("normal","bootstrap"), bs.its = 100,
75 
+                           seed = NULL, verbose = TRUE, ...)
76 
+  standardGeneric("lrt"))
77 
+
78 
+
79 
+#' The optimal discovery procedure
80 
+#'
81 
+#' \code{odp} performs the optimal discovery procedure, which is a new
82 
+#' approach for optimally performing many hypothesis tests in a
83 
+#' high-dimensional study. When testing whether a feature is significant, the
84 
+#' optimal discovery procedure uses information across all features when
85 
+#' testing for significance.
86 
+#'
87 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
88 
+#' @param de.fit \code{S4 object}: \code{\linkS4class{deFit}}. Optional.
89 
+#' @param odp.parms \code{list}: parameters for each cluster. See
90 
+#' \code{\link{kl_clust}}.
91 
+#' @param bs.its \code{numeric}: number of null bootstrap iterations. Default
92 
+#' is 100.
93 
+#' @param n.mods \code{integer}: number of clusters used in
94 
+#' \code{\link{kl_clust}}. Default is 50.
95 
+#' @param seed \code{integer}: set the seed value. Default is NULL.
96 
+#' @param verbose \code{boolean}: print iterations for bootstrap method.
97 
+#' Default is TRUE.
98 
+#' @param ... Additional arguments for \code{\link{qvalue}} and
99 
+#' \code{\link{empPvals}}.
100 
+#'
101 
+#'
102 
+#' @details
103 
+#' The full ODP estimator computationally grows quadratically with
104 
+#' respect to the number of genes. This becomes computationally infeasible at
105 
+#' a certain point. Therefore, an alternative method called mODP is used which
106 
+#' has been shown to provide results that are very similar. mODP utilizes a
107 
+#' k-means clustering algorithm where genes are assigned to a cluster based on
108 
+#' the Kullback-Leiber distance. Each gene is assigned an module-average
109 
+#' parameter to calculate the ODP score and it reduces the computations time
110 
+#' to linear (See Woo, Leek and Storey 2010). If the number of clusters is equal
111 
+#' to the number of genes then the original ODP is implemented. Depending on
112 
+#' the number of hypothesis tests, this can take some time.
113 
+#'
114 
+#' @return \code{\linkS4class{deSet}} object
115 
+#'
116 
+#' @examples
117 
+#' # import data
118 
+#' library(splines)
119 
+#' data(kidney)
120 
+#' age <- kidney$age
121 
+#' sex <- kidney$sex
122 
+#' kidexpr <- kidney$kidexpr
123 
+#' cov <- data.frame(sex = sex, age = age)
124 
+#'
125 
+#' # create models
126 
+#' null_model <- ~sex
127 
+#' full_model <- ~sex + ns(age, df = 4)
128 
+#'
129 
+#' # create deSet object from data
130 
+#' de_obj <- build_models(data = kidexpr, cov = cov,
131 
+#' null.model = null_model, full.model = full_model)
132 
+#'
133 
+#' # odp method
134 
+#' de_odp <- odp(de_obj, bs.its = 30)
135 
+#'
136 
+#' # input a deFit object or ODP parameters ... not necessary
137 
+#' de_fit <- fit_models(de_obj, stat.type = "odp")
138 
+#' de_clust <- kl_clust(de_obj, n.mods = 10)
139 
+#' de_odp <- odp(de_obj, de.fit = de_fit, odp.parms = de_clust,
140 
+#' bs.its = 30)
141 
+#'
142 
+#' # summarize object
143 
+#' summary(de_odp)
144 
+#'
145 
+#' @references
146 
+#' Storey JD. (2007) The optimal discovery procedure: A new approach to
147 
+#' simultaneous significance testing. Journal of the Royal Statistical
148 
+#' Society, Series B, 69: 347-368.
149 
+#'
150 
+#' Storey JD, Dai JY, and Leek JT. (2007) The optimal discovery procedure for
151 
+#' large-scale significance testing, with applications to comparative
152 
+#' microarray experiments. Biostatistics, 8: 414-432.
153 
+#'
154 
+#' Woo S, Leek JT, Storey JD (2010) A computationally efficient modular
155 
+#' optimal discovery procedure. Bioinformatics, 27(4): 509-515.
156 
+#'
157 
+#' @author John Storey, Jeffrey Leek, Andrew Bass
158 
+#'
159 
+#' @seealso \code{\link{kl_clust}}, \code{\link{build_models}} and
160 
+#' \code{\linkS4class{deSet}}
161 
+#'
162 
+#' @export
163 
+setGeneric("odp", function(object, de.fit, odp.parms = NULL, bs.its = 100,
164 
+                           n.mods = 50, seed = NULL, verbose = TRUE, ...)
165 
+  standardGeneric("odp"))
166 
+
167 
+
168 
+#' Modular optimal discovery procedure (mODP)
169 
+#'
170 
+#' \code{kl_clust} is an implementation of mODP that assigns genes to modules
171 
+#' based off of the Kullback-Leibler distance.
172 
+#'
173 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}.
174 
+#' @param de.fit \code{S4 object}: \code{\linkS4class{deFit}}.
175 
+#' @param n.mods \code{integer}: number of clusters.
176 
+#'
177 
+#' @details mODP utilizes a k-means clustering algorithm where genes are
178 
+#' assigned to a cluster based on the Kullback-Leiber distance. Each gene is
179 
+#' assigned an module-average parameter to calculate the ODP score (See Woo,
180 
+#' Leek and Storey 2010 for more details). The mODP and full ODP produce near
181 
+#' exact results but mODP has the advantage of being computationally
182 
+#' feasible.
183 
+#'
184 
+#' @note The results are generally insensitive to the number of modules after a
185 
+#' certain threshold of about n.mods>=50. It is recommended that users
186 
+#' experiment with the number of clusters. If the number of clusters is equal
187 
+#' to the number of genes then the original ODP is implemented. Depending on
188 
+#' the number of hypothesis tests, this can take some time.
189 
+#'
190 
+#' @return
191 
+#' A list with the following slots:
192 
+#' \itemize{
193 
+#'   \item {mu.full: cluster means from full model.}
194 
+#'   \item {mu.null: cluster means from null model.}
195 
+#'   \item {sig.full: cluster standard deviations from full model.}
196 
+#'   \item {sig.null: cluster standard deviations from null model.}
197 
+#'   \item {n.per.mod: total members in each cluster.}
198 
+#'   \item {clustMembers: cluster membership for each gene.}
199 
+#' }
200 
+#'
201 
+#' @examples
202 
+#' # import data
203 
+#' library(splines)
204 
+#' data(kidney)
205 
+#' age <- kidney$age
206 
+#' sex <- kidney$sex
207 
+#' kidexpr <- kidney$kidexpr
208 
+#' cov <- data.frame(sex = sex, age = age)
209 
+#'
210 
+#' # create models
211 
+#' null_model <- ~sex
212 
+#' full_model <- ~sex + ns(age, df = 4)
213 
+#'
214 
+#' # create deSet object from data
215 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
216 
+#' full.model = full_model)
217 
+#'
218 
+#' # mODP method
219 
+#' de_clust <- kl_clust(de_obj)
220 
+#'
221 
+#' # change the number of clusters
222 
+#' de_clust <- kl_clust(de_obj, n.mods = 10)
223 
+#'
224 
+#' # input a deFit object
225 
+#' de_fit <- fit_models(de_obj, stat.type = "odp")
226 
+#' de_clust <- kl_clust(de_obj, de.fit = de_fit)
227 
+#'
228 
+#' @references
229 
+#' Storey JD. (2007) The optimal discovery procedure: A new approach to
230 
+#' simultaneous significance testing. Journal of the Royal Statistical
231 
+#' Society, Series B, 69: 347-368.
232 
+#'
233 
+#' Storey JD, Dai JY, and Leek JT. (2007) The optimal discovery procedure for
234 
+#' large-scale significance testing, with applications to comparative
235 
+#' microarray experiments. Biostatistics, 8: 414-432.
236 
+#'
237 
+#' Woo S, Leek JT, Storey JD (2010) A computationally efficient modular optimal
238 
+#'  discovery procedure. Bioinformatics, 27(4): 509-515.
239 
+#'
240 
+#' @author John Storey, Jeffrey Leek
241 
+#'
242 
+#' @seealso \code{\link{odp}}, \code{\link{lrt}} and
243 
+#' \code{\link{fit_models}}
244 
+#'
245 
+#' @exportMethod kl_clust
246 
+setGeneric("kl_clust", function(object, de.fit = NULL, n.mods = 50)
247 
+  standardGeneric("kl_clust"))
248 
+
249 
+#' Linear regression of the null and full models
250 
+#'
251 
+#' \code{fit_models} fits a linear model to each gene by using the least
252 
+#' squares method. Model fits can be either statistic type "odp" (optimal
253 
+#' discovery procedure) or "lrt" (likelihood ratio test).
254 
+#'
255 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}.
256 
+#' @param stat.type \code{character}: type of statistic to be used. Either
257 
+#' "lrt" or "odp". Default is "lrt".
258 
+#'
259 
+#' @details
260 
+#' If "odp" method is implemented then the null model is removed from the full
261 
+#' model (see Storey 2007).
262 
+#'
263 
+#' @note \code{fit_models} does not have to be called by the user to use
264 
+#' \code{\link{odp}}, \code{\link{lrt}} or \code{\link{kl_clust}} as it is an
265 
+#' optional input and is implemented in the methods. The
266 
+#' \code{\linkS4class{deFit}} object can be created by the user if a different
267 
+#' statistical implementation is required.
268 
+#'
269 
+#' @return \code{\linkS4class{deFit}} object
270 
+#'
271 
+#' @examples
272 
+#' # import data
273 
+#' library(splines)
274 
+#' data(kidney)
275 
+#' age <- kidney$age
276 
+#' sex <- kidney$sex
277 
+#' kidexpr <- kidney$kidexpr
278 
+#' cov <- data.frame(sex = sex, age = age)
279 
+#'
280 
+#' # create models
281 
+#' null_model <- ~sex
282 
+#' full_model <- ~sex + ns(age, df = 4)
283 
+#'
284 
+#' # create deSet object from data
285 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
286 
+#' full.model = full_model)
287 
+#'
288 
+#' # retrieve statistics from linear regression for each gene
289 
+#' fit_lrt <- fit_models(de_obj, stat.type = "lrt") # lrt method
290 
+#' fit_odp <- fit_models(de_obj, stat.type = "odp") # odp method
291 
+#'
292 
+#' # summarize object
293 
+#' summary(fit_odp)
294 
+#'
295 
+#' @references
296 
+#' Storey JD. (2007) The optimal discovery procedure: A new approach to
297 
+#' simultaneous significance testing. Journal of the Royal Statistical
298 
+#' Society, Series B, 69: 347-368.
299 
+#'
300 
+#' Storey JD, Dai JY, and Leek JT. (2007) The optimal discovery procedure for
301 
+#' large-scale significance testing, with applications to comparative
302 
+#' microarray experiments. Biostatistics, 8: 414-432.
303 
+#'
304 
+#' Storey JD, Xiao W, Leek JT, Tompkins RG, and Davis RW. (2005) Significance
305 
+#' analysis of time course microarray experiments. Proceedings of the National
306 
+#' Academy of Sciences, 102: 12837-12842.
307 
+#'
308 
+#' @seealso \code{\linkS4class{deFit}}, \code{\link{odp}} and
309 
+#' \code{\link{lrt}}
310 
+#'
311 
+#' @author John Storey
312 
+#' @exportMethod fit_models
313 
+setGeneric("fit_models",
314 
+           function(object, stat.type = c("lrt", "odp")) {
315 
+             standardGeneric("fit_models")
316 
+           })
317 
+
318 
+#' Create a deSet object from an ExpressionSet
319 
+#'
320 
+#' Creates a \code{\linkS4class{deSet}} object that extends the
321 
+#' \code{\link{ExpressionSet}} object.
322 
+#'
323 
+#' @param object \code{S4 object}: \code{\link{ExpressionSet}}
324 
+#' @param full.model \code{formula}: full model containing the both the
325 
+#' adjustment and the biological variables for the experiment.
326 
+#' @param null.model \code{formula}: null model containing the adjustment
327 
+#' variables for the experiment.
328 
+#' @param individual \code{factor}: information on repeated samples in
329 
+#' experiment.
330 
+#'
331 
+#' @note It is essential that the null and full models have the same variables
332 
+#' as the ExpressionSet phenoType column names.
333 
+#'
334 
+#' @return \code{\linkS4class{deSet}} object
335 
+#'
336 
+#' @examples
337 
+#' # import data
338 
+#' library(splines)
339 
+#' data(kidney)
340 
+#' age <- kidney$age
341 
+#' sex <- kidney$sex
342 
+#' kidexpr <- kidney$kidexpr
343 
+#' cov <- data.frame(sex = sex, age = age)
344 
+#' pDat <- as(cov, "AnnotatedDataFrame")
345 
+#' exp_set <- ExpressionSet(assayData = kidexpr, phenoData = pDat)
346 
+#'
347 
+#' # create models
348 
+#' null_model <- ~sex
349 
+#' full_model <- ~sex + ns(age, df = 4)
350 
+#'
351 
+#' # create deSet object from data
352 
+#' de_obj <- deSet(exp_set, null.model = null_model,
353 
+#' full.model = full_model)
354 
+#'
355 
+#' # optionally add individuals to experiment, in this case there are 36
356 
+#' # individuals that were sampled twice
357 
+#' indSamples <- as.factor(rep(1:36, each = 2))
358 
+#' de_obj <- deSet(exp_set, null.model = null_model,
359 
+#' full.model = full_model, ind = indSamples)
360 
+#' summary(de_obj)
361 
+#' @seealso \code{\linkS4class{deSet}}, \code{\link{odp}} and
362 
+#' \code{\link{lrt}}
363 
+#'
364 
+#' @author John Storey, Andrew Bass
365 
+#'
366 
+#' @export
367 
+setGeneric("deSet", function(object, full.model, null.model,
368 
+                             individual=NULL) standardGeneric("deSet"))
369 
+
370 
+#' Estimate the q-values for a given set of p-values
371 
+#'
372 
+#' Runs \code{\link{qvalue}} on a \code{\linkS4class{deSet}} object.
373 
+#'
374 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
375 
+#' @param ... Additional arguments for \code{\link{qvalue}}
376 
+#'
377 
+#' @return \code{\linkS4class{deSet}} object
378 
+#'
379 
+#' @examples
380 
+#' # import data
381 
+#' library(splines)
382 
+#' data(kidney)
383 
+#' age <- kidney$age
384 
+#' sex <- kidney$sex
385 
+#' kidexpr <- kidney$kidexpr
386 
+#' cov <- data.frame(sex = sex, age = age)
387 
+#'
388 
+#' # create models
389 
+#' null_model <- ~sex
390 
+#' full_model <- ~sex + ns(age, df = 4)
391 
+#'
392 
+#' # create deSet object from data
393 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
394 
+#' full.model = full_model)
395 
+#'
396 
+#' # Run lrt (or odp) and apply_qvalue
397 
+#' de_lrt <- lrt(de_obj)
398 
+#' de_lrt <- apply_qvalue(de_lrt, fdr.level = 0.05,
399 
+#' pi0.method = "bootstrap", adj=1.2)
400 
+#' summary(de_lrt)
401 
+#'
402 
+#' @references
403 
+#' Storey JD and Tibshirani R. (2003) Statistical significance for
404 
+#' genome-wide studies. Proceedings of the National Academy of Sciences,
405 
+#' 100: 9440-9445
406 
+#'
407 
+#' @seealso \code{\linkS4class{deSet}}, \code{\link{odp}} and
408 
+#' \code{\link{lrt}}
409 
+#'
410 
+#' @author John Storey, Andrew Bass
411 
+#'
412 
+#' @export
413 
+setGeneric("apply_qvalue", function(object, ...)
414 
+  standardGeneric("apply_qvalue"))
415 
+
416 
+#' Estimate surrogate variables
417 
+#'
418 
+#' Runs \code{\link{sva}} on the null and full models in
419 
+#' \code{\linkS4class{deSet}}. See \code{\link{sva}} for additional details.
420 
+#'
421 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
422 
+#' @param ... Additional arguments for \code{\link{sva}}
423 
+#'
424 
+#' @return \code{\linkS4class{deSet}} object
425 
+#'
426 
+#' @examples
427 
+#' # import data
428 
+#' library(splines)
429 
+#' data(kidney)
430 
+#' age <- kidney$age
431 
+#' sex <- kidney$sex
432 
+#' kidexpr <- kidney$kidexpr
433 
+#' cov <- data.frame(sex = sex, age = age)
434 
+#'
435 
+#' # create models
436 
+#' null_model <- ~sex
437 
+#' full_model <- ~sex + ns(age, df = 4)
438 
+#'
439 
+#' # create deSet object from data
440 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
441 
+#' full.model = full_model)
442 
+#'
443 
+#' # run surrogate variable analysis
444 
+#' de_sva <- apply_sva(de_obj)
445 
+#'
446 
+#' # run odp/lrt with surrogate variables added
447 
+#' de_odp <- odp(de_sva, bs.its = 30)
448 
+#' summary(de_odp)
449 
+#' @seealso \code{\linkS4class{deSet}}, \code{\link{odp}} and
450 
+#' \code{\link{lrt}}
451 
+#'
452 
+#' @references
453 
+#' Leek JT, Storey JD (2007) Capturing Heterogeneity in Gene Expression
454 
+#' Studies by Surrogate Variable Analysis. PLoS Genet 3(9): e161.
455 
+#' doi:10.1371/journal.pgen.0030161
456 
+#'
457 
+#' @author John Storey, Jeffrey Leek, Andrew Bass
458 
+#' @export
459 
+setGeneric("apply_sva", function(object, ...)
460 
+  standardGeneric("apply_sva"))
461 
+
462 
+#' Supervised normalization of data in edge
463 
+#'
464 
+#' Runs \code{snm} on a deSet object based on the null and full models in
465 
+#' \code{\linkS4class{deSet}}. See \code{\link{snm}} for additional details
466 
+#' on the algorithm.
467 
+#'
468 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
469 
+#' @param int.var \code{data frame}: intensity-dependent effects.
470 
+#' @param ... Additional arguments for \code{\link{snm}}
471 
+#'
472 
+#' @return \code{apply_snm} returns an \code{\linkS4class{deSet}} object.
473 
+#'
474 
+#' @references
475 
+#' Mechan BH, Nelson PS, Storey JD. Supervised normalization of microarrays.
476 
+#' Bioinformatics 2010;26:1308-15
477 
+#'
478 
+#' @examples
479 
+#' # simulate data
480 
+#' library(snm)
481 
+#' singleChannel <- sim.singleChannel(12345)
482 
+#' data <- singleChannel$raw.data
483 
+#'
484 
+#' # create deSet object using build_models (can use ExpressionSet see manual)
485 
+#' cov <- data.frame(grp = singleChannel$bio.var[,2])
486 
+#' full_model <- ~grp
487 
+#' null_model <- ~1
488 
+#'
489 
+#' # create deSet object using build_models
490 
+#' de_obj <- build_models(data = data, cov = cov, full.model = full_model,
491 
+#' null.model = null_model)
492 
+#'
493 
+#' # run snm using intensity-dependent adjustment variable
494 
+#' de_snm <- apply_snm(de_obj, int.var = singleChannel$int.var,
495 
+#' verbose = FALSE, num.iter = 1)
496 
+#'
497 
+#' @seealso \code{\linkS4class{deSet}}, \code{\link{odp}} and
498 
+#' \code{\link{lrt}}
499 
+#'
500 
+#' @author John Storey, Andrew Bass
501 
+#' @export
502 
+setGeneric("apply_snm", function(object, int.var, ...)
503 
+  standardGeneric("apply_snm"))
504 
+
505 
+#' Full model equation
506 
+#'
507 
+#' These generic functions access and set the full model for
508 
+#' \code{\linkS4class{deSet}} object.
509 
+#'
510 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
511 
+#' @param value \code{formula}: The experiment design for the full model.
512 
+#'
513 
+#' @examples
514 
+#' # import data
515 
+#' library(splines)
516 
+#' data(kidney)
517 
+#' age <- kidney$age
518 
+#' sex <- kidney$sex
519 
+#' kidexpr <- kidney$kidexpr
520 
+#' cov <- data.frame(sex = sex, age = age)
521 
+#'
522 
+#' # create models
523 
+#' null_model <- ~sex
524 
+#' full_model <- ~sex + ns(age, df = 4)
525 
+#'
526 
+#' # create deSet object from data
527 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
528 
+#' full.model = full_model)
529 
+#'
530 
+#' # extract out the full model equation
531 
+#' mod_full <- fullModel(de_obj)
532 
+#'
533 
+#' # change the full model in the experiment
534 
+#' fullModel(de_obj) <- ~sex + ns(age, df = 2)
535 
+#'
536 
+#'
537 
+#' @return the formula for the full model.
538 
+#'
539 
+#' @author John Storey, Andrew Bass
540 
+#'
541 
+#' @seealso \code{\linkS4class{deSet}}
542 
+#'
543 
+#' @export
544 
+setGeneric("fullModel", function(object) standardGeneric("fullModel"))
545 
+
546 
+#' @rdname fullModel
547 
+#' @export
548 
+setGeneric("fullModel<-", function(object, value) {
549 
+  standardGeneric("fullModel<-")
550 
+})
551 
+
552 
+#' Null model equation from deSet object
553 
+#'
554 
+#' These generic functions access and set the null model for
555 
+#' \code{\linkS4class{deSet}} object.
556 
+#'
557 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
558 
+#' @param value \code{formula}: The experiment design for the null model.
559 
+#'
560 
+#' @return \code{nullModel} returns the formula for the null model.
561 
+#'
562 
+#' @examples
563 
+#' # import data
564 
+#' library(splines)
565 
+#' data(kidney)
566 
+#' age <- kidney$age
567 
+#' sex <- kidney$sex
568 
+#' kidexpr <- kidney$kidexpr
569 
+#' cov <- data.frame(sex = sex, age = age)
570 
+#'
571 
+#' # create models
572 
+#' null_model <- ~sex
573 
+#' full_model <- ~sex + ns(age, df = 4)
574 
+#'
575 
+#' # create deSet object from data
576 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
577 
+#' full.model = full_model)
578 
+#'
579 
+#' # extract the null model equation
580 
+#' mod_null <- nullModel(de_obj)
581 
+#'
582 
+#' # change null model in experiment but must update full model
583 
+#' nullModel(de_obj) <- ~1
584 
+#' fullModel(de_obj) <- ~1 + ns(age, df=4)
585 
+#' @author John Storey, Andrew Bass
586 
+#'
587 
+#' @seealso \code{\linkS4class{deSet}}
588 
+#'
589 
+#' @keywords nullModel, nullModel<-
590 
+#'
591 
+#' @exportMethod nullModel
592 
+setGeneric("nullModel", function(object) standardGeneric("nullModel"))
593 
+
594 
+#' @rdname nullModel
595 
+#' @export
596 
+setGeneric("nullModel<-", function(object, value) {
597 
+  standardGeneric("nullModel<-")
598 
+})
599 
+
600 
+#' Matrix representation of null model
601 
+#'
602 
+#' These generic functions access and set the null matrix for
603 
+#' \code{\linkS4class{deSet}} object.
604 
+#'
605 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
606 
+#' @param value \code{matrix}: null model matrix where columns are covariates
607 
+#' and rows are observations
608 
+#'
609 
+#' @return \code{nullMatrix} returns the value of the null model matrix.
610 
+#'
611 
+#' @examples
612 
+#' # import data
613 
+#' library(splines)
614 
+#' data(kidney)
615 
+#' age <- kidney$age
616 
+#' sex <- kidney$sex
617 
+#' kidexpr <- kidney$kidexpr
618 
+#' cov <- data.frame(sex = sex, age = age)
619 
+#'
620 
+#' # create models
621 
+#' null_model <- ~sex
622 
+#' full_model <- ~sex + ns(age, df = 4)
623 
+#'
624 
+#' # create deSet object from data
625 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
626 
+#' full.model = full_model)
627 
+#'
628 
+#' # extract the null model as a matrix
629 
+#' mat_null <- nullMatrix(de_obj)
630 
+#'
631 
+#' @author John Storey, Andrew Bass
632 
+#'
633 
+#' @seealso \code{\linkS4class{deSet}}, \code{\link{fullModel}} and
634 
+#' \code{\link{fullModel}}
635 
+#'
636 
+#' @export
637 
+setGeneric("nullMatrix", function(object) standardGeneric("nullMatrix"))
638 
+
639 
+#' @rdname nullMatrix
640 
+#' @export
641 
+setGeneric("nullMatrix<-", function(object, value) {
642 
+  standardGeneric("nullMatrix<-")
643 
+})
644 
+
645 
+#' Matrix representation of full model
646 
+#'
647 
+#' These generic functions access and set the full matrix for
648 
+#' \code{\linkS4class{deSet}} object.
649 
+#'
650 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
651 
+#' @param value \code{matrix}: full model matrix where the columns are the
652 
+#' covariates and rows are observations
653 
+#'
654 
+#' @return \code{fullMatrix} returns the value of the full model matrix.
655 
+#'
656 
+#' @examples
657 
+#' # import data
658 
+#' library(splines)
659 
+#' data(kidney)
660 
+#' age <- kidney$age
661 
+#' sex <- kidney$sex
662 
+#' kidexpr <- kidney$kidexpr
663 
+#' cov <- data.frame(sex = sex, age = age)
664 
+#'
665 
+#' # create models
666 
+#' null_model <- ~sex
667 
+#' full_model <- ~sex + ns(age, df = 4)
668 
+#'
669 
+#' # create deSet object from data
670 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
671 
+#' full.model = full_model)
672 
+#'
673 
+#' # extract the full model equation as a matrix
674 
+#' mat_full <- fullMatrix(de_obj)
675 
+#' @author Andrew Bass, John Storey
676 
+#'
677 
+#' @seealso \code{\linkS4class{deSet}}, \code{\link{fullModel}}
678 
+#'
679 
+#' @export
680 
+setGeneric("fullMatrix", function(object) standardGeneric("fullMatrix"))
681 
+
682 
+#' @rdname fullMatrix
683 
+#' @export
684 
+setGeneric("fullMatrix<-", function(object, value) {
685 
+  standardGeneric("fullMatrix<-")
686 
+})
687 
+
688 
+
689 
+#' Access/set qvalue slot
690 
+#'
691 
+#' These generic functions access and set the \code{qvalue} object in the
692 
+#' \code{\linkS4class{deSet}} object.
693 
+#'
694 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
695 
+#' @param value S3 \code{object}: \code{\link{qvalue}}
696 
+#'
697 
+#' @return  \code{qvalueObj} returns a \code{\link{qvalue}} object.
698 
+#'
699 
+#' @examples
700 
+#' # import data
701 
+#' library(splines)
702 
+#' library(qvalue)
703 
+#' data(kidney)
704 
+#' age <- kidney$age
705 
+#' sex <- kidney$sex
706 
+#' kidexpr <- kidney$kidexpr
707 
+#' cov <- data.frame(sex = sex, age = age)
708 
+#'
709 
+#' # create models
710 
+#' null_model <- ~sex
711 
+#' full_model <- ~sex + ns(age, df = 4)
712 
+#'
713 
+#' # create deSet object from data
714 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
715 
+#' full.model = full_model)
716 
+#'
717 
+#' # run the odp method
718 
+#' de_odp <- odp(de_obj, bs.its = 20)
719 
+#'
720 
+#' # extract out significance results
721 
+#' qval_obj <- qvalueObj(de_odp)
722 
+#'
723 
+#' # run qvalue and assign it to deSet slot
724 
+#' pvals <- qval_obj$pvalues
725 
+#' qval_new <- qvalue(pvals, pfdr = TRUE, fdr.level = 0.1)
726 
+#' qvalueObj(de_odp) <- qval_new
727 
+#'
728 
+#' @author John Storey, Andrew Bass
729 
+#'
730 
+#' @seealso \code{\link{lrt}}, \code{\link{odp}} and
731 
+#' \code{\linkS4class{deSet}}
732 
+#'
733 
+#' @export
734 
+setGeneric("qvalueObj", function(object) standardGeneric("qvalueObj"))
735 
+
736 
+#' @rdname qvalueObj
737 
+#' @export
738 
+setGeneric("qvalueObj<-", function(object, value) {
739 
+  standardGeneric("qvalueObj<-")
740 
+})
741 
+
742 
+#' Individuals sampled in experiment
743 
+#'
744 
+#' These generic functions access and set the individual slot in
745 
+#' \code{\linkS4class{deSet}}.
746 
+#'
747 
+#' @param object \code{\linkS4class{deSet}}
748 
+#' @param value \code{factor}: identifier for each observation. Important
749 
+#' if the same individuals are sampled multiple times.
750 
+#'
751 
+#' @return \code{individual} returns information regarding individuals
752 
+#' in the experiment.
753 
+#'
754 
+#' @examples
755 
+#' library(splines)
756 
+#' # import data
757 
+#' data(endotoxin)
758 
+#' ind <- endotoxin$ind
759 
+#' time <- endotoxin$t
760 
+#' class <- endotoxin$class
761 
+#' endoexpr <- endotoxin$endoexpr
762 
+#' cov <- data.frame(individual = ind, time = time, class = class)
763 
+#'
764 
+#' # create ExpressionSet object
765 
+#' pDat <- as(cov, "AnnotatedDataFrame")
766 
+#' exp_set <- ExpressionSet(assayData = endoexpr, phenoData = pDat)
767 
+#'
768 
+#' # formulate null and full models in experiement
769 
+#' # note: interaction term is a way of taking into account group effects
770 
+#' mNull <- ~ns(time, df=4, intercept = FALSE)
771 
+#' mFull <- ~ns(time, df=4, intercept = FALSE) +
772 
+#' ns(time, df=4, intercept = FALSE):class + class
773 
+#'
774 
+#' # create deSet object
775 
+#' de_obj <- deSet(exp_set, full.model = mFull, null.model = mNull,
776 
+#' individual = ind)
777 
+#'
778 
+#' # extract out the individuals factor
779 
+#' ind_exp <- individual(de_obj)
780 
+#'
781 
+#' @author John Storey, Andrew Bass
782 
+#'
783 
+#' @seealso \code{\linkS4class{deSet}}
784 
+#'
785 
+#' @export
786 
+setGeneric("individual", function(object) standardGeneric("individual"))
787 
+
788 
+#' @rdname individual
789 
+#' @export
790 
+setGeneric("individual<-", function(object, value) {
791 
+  standardGeneric("individual<-")
792 
+})
793 
+
794 
+#' Regression coefficients from full model fit
795 
+#'
796 
+#' Access the full model fitted coefficients of a
797 
+#' \code{\linkS4class{deFit}} object.
798 
+#'
799 
+#' @param object \code{S4 object}: \code{\linkS4class{deFit}}
800 
+#'
801 
+#' @return \code{betaCoef} returns the regression coefficients.
802 
+#'
803 
+#' @author John Storey, Andrew Bass
804 
+#'
805 
+#' @seealso \code{\link{fit_models}}
806 
+#'
807 
+#' @examples
808 
+#' # import data
809 
+#' library(splines)
810 
+#' data(kidney)
811 
+#' age <- kidney$age
812 
+#' sex <- kidney$sex
813 
+#' kidexpr <- kidney$kidexpr
814 
+#' cov <- data.frame(sex = sex, age = age)
815 
+#'
816 
+#' # create models
817 
+#' null_model <- ~sex
818 
+#' full_model <- ~sex + ns(age, df = 4)
819 
+#'
820 
+#' # create deSet object from data
821 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
822 
+#' full.model = full_model)
823 
+#'
824 
+#' # run fit_models to get model fits
825 
+#' de_fit <- fit_models(de_obj)
826 
+#'
827 
+#' # extract beta coefficients
828 
+#' beta <- betaCoef(de_fit)
829 
+#'
830 
+#' @export
831 
+setGeneric("betaCoef", function(object) standardGeneric("betaCoef"))
832 
+
833 
+#' Statistical method used in analysis
834 
+#'
835 
+#' Access the statistic type in a \code{\linkS4class{deFit}} object. Can
836 
+#' either be the optimal discovery procedure (odp) or the likelihood ratio
837 
+#' test (lrt).
838 
+#'
839 
+#' @param object \code{S4 object}: \code{\linkS4class{deFit}}
840 
+#'
841 
+#' @examples
842 
+#' # import data
843 
+#' library(splines)
844 
+#' data(kidney)
845 
+#' age <- kidney$age
846 
+#' sex <- kidney$sex
847 
+#' kidexpr <- kidney$kidexpr
848 
+#' cov <- data.frame(sex = sex, age = age)
849 
+#'
850 
+#' # create models
851 
+#' null_model <- ~sex
852 
+#' full_model <- ~sex + ns(age, df = 4)
853 
+#'
854 
+#' # create deSet object from data
855 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
856 
+#' full.model = full_model)
857 
+#'
858 
+#' # run fit_models to get model fits
859 
+#' de_fit <- fit_models(de_obj)
860 
+#'
861 
+#' # extract the statistic type of model fits
862 
+#' stat_type <- sType(de_fit)
863 
+#'
864 
+#' @return \code{sType} returns the statistic type- either "odp" or "lrt".
865 
+#'
866 
+#' @author John Storey, Andrew Bass
867 
+#'
868 
+#' @seealso \code{\link{fit_models}}, \code{\linkS4class{deFit}} and
869 
+#' \code{\linkS4class{deSet}}
870 
+#'
871 
+#' @keywords sType
872 
+#'
873 
+#' @exportMethod sType
874 
+setGeneric("sType", function(object) standardGeneric("sType"))
875 
+
876 
+#' Fitted data from the full model
877 
+#'
878 
+#' Access the fitted data from the full model in a
879 
+#' \code{\linkS4class{deFit}} object.
880 
+#'
881 
+#' @param object \code{S4 object}: \code{\linkS4class{deFit}}
882 
+#'
883 
+#' @usage fitFull(object)
884 
+#'
885 
+#' @return \code{fitFull} returns a matrix of fitted values from full model.
886 
+#'
887 
+#' @examples
888 
+#' # import data
889 
+#' library(splines)
890 
+#' data(kidney)
891 
+#' age <- kidney$age
892 
+#' sex <- kidney$sex
893 
+#' kidexpr <- kidney$kidexpr
894 
+#' cov <- data.frame(sex = sex, age = age)
895 
+#'
896 
+#' # create models
897 
+#' null_model <- ~sex
898 
+#' full_model <- ~sex + ns(age, df = 4)
899 
+#'
900 
+#' # create deSet object from data
901 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
902 
+#' full.model = full_model)
903 
+#'
904 
+#' # run fit_models to get model fits
905 
+#' de_fit <- fit_models(de_obj)
906 
+#'
907 
+#' # extract fitted values for full model
908 
+#' fitted_full <- fitFull(de_fit)
909 
+#'
910 
+#' @author John Storey, Andrew Bass
911 
+#'
912 
+#' @seealso \code{\link{fit_models}}
913 
+#'
914 
+#' @export
915 
+setGeneric("fitFull", function(object) standardGeneric("fitFull"))
916 
+
917 
+#' Fitted data from the null model
918 
+#'
919 
+#' Access the fitted data from the null model in an
920 
+#' \code{\linkS4class{deFit}} object.
921 
+#'
922 
+#' @param object \code{S4 object}: \code{\linkS4class{deFit}}
923 
+#'
924 
+#' @usage fitNull(object)
925 
+#'
926 
+#' @return \code{fitNull} returns a matrix of fitted values from null model.
927 
+#'
928 
+#' @examples
929 
+#' # import data
930 
+#' library(splines)
931 
+#' data(kidney)
932 
+#' age <- kidney$age
933 
+#' sex <- kidney$sex
934 
+#' kidexpr <- kidney$kidexpr
935 
+#' cov <- data.frame(sex = sex, age = age)
936 
+#'
937 
+#' # create models
938 
+#' null_model <- ~sex
939 
+#' full_model <- ~sex + ns(age, df = 4)
940 
+#'
941 
+#' # create deSet object from data
942 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
943 
+#' full.model = full_model)
944 
+#'
945 
+#' # run fit_models to get model fits
946 
+#' de_fit <- fit_models(de_obj)
947 
+#'
948 
+#' # extract fitted values from null model
949 
+#' fitted_null <- fitNull(de_fit)
950 
+#'
951 
+#' @author  John Storey, Andrew Bass
952 
+#'
953 
+#' @seealso \code{\link{fit_models}}
954 
+#'
955 
+#' @export
956 
+setGeneric("fitNull", function(object) standardGeneric("fitNull"))
957 
+
958 
+#' Residuals of full model fit
959 
+#'
960 
+#' Access the fitted full model residuals in an \code{\linkS4class{deFit}}
961 
+#' object.
962 
+#'
963 
+#' @param object \code{S4 object}: \code{\linkS4class{deFit}}
964 
+#'
965 
+#' @usage resFull(object)
966 
+#'
967 
+#' @return \code{resFull} returns a matrix of residuals from full model.
968 
+#'
969 
+#' @examples
970 
+#' # import data
971 
+#' library(splines)
972 
+#' data(kidney)
973 
+#' age <- kidney$age
974 
+#' sex <- kidney$sex
975 
+#' kidexpr <- kidney$kidexpr
976 
+#' cov <- data.frame(sex = sex, age = age)
977 
+#'
978 
+#' # create models
979 
+#' null_model <- ~sex
980 
+#' full_model <- ~sex + ns(age, df = 4)
981 
+#'
982 
+#' # create deSet object from data
983 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
984 
+#' full.model = full_model)
985 
+#'
986 
+#' # run fit_models to get model fits
987 
+#' de_fit <- fit_models(de_obj)
988 
+#'
989 
+#' # extract out the full residuals from the model fit
990 
+#' res_full <- resFull(de_fit)
991 
+#'
992 
+#' @author John Storey, Andrew Bass
993 
+#'
994 
+#' @seealso \code{\link{fit_models}}
995 
+#'
996 
+#' @export
997 
+setGeneric("resFull", function(object) standardGeneric("resFull"))
998 
+
999 
+#' Residuals of null model fit
1000 
+#'
1001 
+#' Access the fitted null model residuals in an \code{\linkS4class{deFit}}
1002 
+#' object.
1003 
+#'
1004 
+#' @param object \code{S4 object}: \code{\linkS4class{deFit}}
1005 
+#'
1006 
+#' @usage resNull(object)
1007 
+#'
1008 
+#' @return \code{resNull} returns a matrix of residuals from null model.
1009 
+#'
1010 
+#' @examples
1011 
+#' # import data
1012 
+#' library(splines)
1013 
+#' data(kidney)
1014 
+#' age <- kidney$age
1015 
+#' sex <- kidney$sex
1016 
+#' kidexpr <- kidney$kidexpr
1017 
+#' cov <- data.frame(sex = sex, age = age)
1018 
+#'
1019 
+#' # create models
1020 
+#' null_model <- ~sex
1021 
+#' full_model <- ~sex + ns(age, df = 4)
1022 
+#'
1023 
+#' # create deSet object from data
1024 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
1025 
+#' full.model = full_model)
1026 
+#'
1027 
+#' # run fit_models to get model fits
1028 
+#' de_fit <- fit_models(de_obj)
1029 
+#'
1030 
+#' # extract out the null residuals from the model fits
1031 
+#' res_null <- resNull(de_fit)
1032 
+#' @author John Storey, Andrew Bass
1033 
+#'
1034 
+#' @seealso  \code{\link{fit_models}}
1035 
+#'
1036 
+#' @export
1037 
+setGeneric("resNull", function(object) standardGeneric("resNull"))
1038 
+
1039 
+#' Summary of deFit and deSet
1040 
+#'
1041 
+#' Summary of \code{\linkS4class{deFit}} and \code{\linkS4class{deSet}} objects.
1042 
+#'
1043 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
1044 
+#' @param \dots additional parameters
1045 
+#'
1046 
+#' @examples
1047 
+#' # import data
1048 
+#' library(splines)
1049 
+#' data(kidney)
1050 
+#' age <- kidney$age
1051 
+#' sex <- kidney$sex
1052 
+#' kidexpr <- kidney$kidexpr
1053 
+#' cov <- data.frame(sex = sex, age = age)
1054 
+#'
1055 
+#' # create models
1056 
+#' null_model <- ~sex
1057 
+#' full_model <- ~sex + ns(age, df = 4)
1058 
+#'
1059 
+#' # create deSet object from data
1060 
+#' de_obj <- build_models(data = kidexpr, cov = cov, null.model = null_model,
1061 
+#' full.model = full_model)
1062 
+#'
1063 
+#' # get summary
1064 
+#' summary(de_obj)
1065 
+#'
1066 
+#' # run odp and summarize
1067 
+#' de_odp <- odp(de_obj, bs.its= 20)
1068 
+#' summary(de_odp)
1069 
+#' @author John Storey, Andrew Bass
1070 
+#'
1071 
+#' @return
1072 
+#' Summary of \code{\linkS4class{deSet}} object
1073 
+#'
1074 
+#' @keywords summary
1075 
+#'
1076 
+#' @export summary
1077 
+setGeneric("summary")
1078 
+
1079 
+#' Show function for deFit and deSet
1080 
+#'
1081 
+#' Show function for \code{\linkS4class{deFit}} and \code{\linkS4class{deSet}}
1082 
+#' objects.
1083 
+#'
1084 
+#' @param object \code{S4 object}: \code{\linkS4class{deSet}}
1085 
+#' @param \dots additional parameters
1086 
+#'
1087 
+#' @examples
1088 
+#' # import data
1089 
+#' library(splines)
1090 
+#' data(kidney)
1091 
+#' age <- kidney$age
1092 
+#' sex <- kidney$sex
1093 
+#' kidexpr <- kidney$kidexpr
1094 
+#' cov <- data.frame(sex = sex, age = age)
1095 
+#'
1096 
+#' # create models