1.0.1 release cleanup

DESCRIPTION

@@ -1,9 +1,9 @@
 Package: topconfects
 Title: Top Results by Confident Effect Size
-Version: 0.0.1
+Version: 1.0.1
 Authors@R: person("Paul", "Harrison", email = "pfh@logarithmic.net", role = c("aut", "cre"))
 Description: Uses limma's treat or edgeR's glmTreat to rank genes (or other
-    features) by effect size.
+    features) by confident log2 fold change.
 Depends:
     R (>= 3.3.0)
 Imports:

NAMESPACE

@@ -5,25 +5,11 @@ export(confects_plot_me)
 export(edger_confects)
 export(edger_group_confects)
 export(effect_contrast)
-export(effect_contrast_ratio)
-export(effect_gamma)
 export(effect_link_log2)
-export(effect_rss)
-export(effect_rssm)
-export(effect_sd)
 export(effect_shift)
 export(effect_shift_log2)
-export(effect_shift_stepdown)
-export(effect_shift_stepdown_log2)
-export(effect_shift_stepup)
-export(effect_shift_stepup_log2)
-export(group_effect_rssm)
 export(group_effect_shift)
 export(group_effect_shift_log2)
-export(group_effect_shift_stepdown)
-export(group_effect_shift_stepdown_log2)
-export(group_effect_shift_stepup)
-export(group_effect_shift_stepup_log2)
 export(limma_confects)
 export(limma_group_confects)
 export(limma_nonlinear_confects)

R/fit.R

@@ -156,21 +156,21 @@ effect_contrast <- function(contrast) {
 }
 
 
-#' Ratio of contrasts effect.
-#'
-#' Ratio of two contrasts, i.e. \code{sum(contrast1*beta)/sum(contrast2*beta)}.
-#'
-#' @param contrast1 First contrast weights, a vector with the same length as the number of columns in the design matrix.
-#'
-#' @param contrast2 Second contrast weights, a vector with the same length as the number of columns in the design matrix.
-#'
-#' \code{sum(contrast2*beta)} should always produce a positive value.
-#'
-#' @return
-#'
-#' An object defining how to calculate an effect size.
-#' 
-#' @export
+# Ratio of contrasts effect.
+#
+# Ratio of two contrasts, i.e. \code{sum(contrast1*beta)/sum(contrast2*beta)}.
+#
+# @param contrast1 First contrast weights, a vector with the same length as the number of columns in the design matrix.
+#
+# @param contrast2 Second contrast weights, a vector with the same length as the number of columns in the design matrix.
+#
+# \code{sum(contrast2*beta)} should always produce a positive value.
+#
+# @return
+#
+# An object defining how to calculate an effect size.
+# 
+# @export
 effect_contrast_ratio <- function(contrast1, contrast2) {
     list(
         signed = TRUE,
@@ -190,19 +190,19 @@ effect_contrast_ratio <- function(contrast1, contrast2) {
 }
 
 
-#' Standard-deviation of a set of coefficients as effect size
-#'
-#' This is intended as the effect size version of an ANOVA. For effect_sd, the effect size is the standard deviation of some coefficients about their mean. For effect_rssm, it is the root sum of squared differences from the mean. For effect_rss, it is simply the square root of the sum of squared coefficients.
-#'
-#' \code{effect_rssm} may be better suited to comparing effect sizes from designs with differing numbers of coefficients, such as differential exon usage.
-#'
-#' @param coef The column numbers of the design matrix for the relevant coefficients.
-#'
-#' @return
-#'
-#' An object defining how to calculate an effect size.
-#'
-#' @export
+# Standard-deviation of a set of coefficients as effect size
+#
+# This is intended as the effect size version of an ANOVA. For effect_sd, the effect size is the standard deviation of some coefficients about their mean. For effect_rssm, it is the root sum of squared differences from the mean. For effect_rss, it is simply the square root of the sum of squared coefficients.
+#
+# \code{effect_rssm} may be better suited to comparing effect sizes from designs with differing numbers of coefficients, such as differential exon usage.
+#
+# @param coef The column numbers of the design matrix for the relevant coefficients.
+#
+# @return
+#
+# An object defining how to calculate an effect size.
+#
+# @export
 effect_sd <- function(coef) {
     n <- length(coef)
     assert_that(n > 1)   #n=2 case may be problematic
@@ -230,8 +230,8 @@ effect_sd <- function(coef) {
 }
 
 
-#' @rdname effect_sd
-#' @export
+# @rdname effect_sd
+# @export
 effect_rssm <- function(coef) {
     n <- length(coef)
     assert_that(n > 1)   #n=2 case may be problematic
@@ -259,8 +259,8 @@ effect_rssm <- function(coef) {
 }
 
 
-#' @rdname effect_sd
-#' @export
+# @rdname effect_sd
+# @export
 effect_rss <- function(coef) {
     list(
         signed = FALSE,
@@ -294,8 +294,8 @@ effect_rss <- function(coef) {
 #'
 #' Note that this effect size is not symmetric: \code{effect_shift_log2(c(1,2),c(3,4))} and \code{effect_shift_log2(c(1,3),c(2,4))} will give different results.
 #'
-#' The _stepdown and _stepup versions are for cumulative distributions. These are most useful in group effect form, where they can be used to examine shifts in start or end of transcriptions from RNA-seq or microarray data.
-#'
+# The _stepdown and _stepup versions are for cumulative distributions. These are most useful in group effect form, where they can be used to examine shifts in start or end of transcriptions from RNA-seq or microarray data.
+#
 #' @param coef1 Column numbers in the design matrix for the first condition, in some meaningful order.
 #'
 #' @param coef2 Corresponding column numbers for the second condition.
@@ -313,8 +313,8 @@ effect_shift <- function(coef1, coef2) {
     effect_shift_inner(n, sign_mat, coef1, coef2, c(-1,1))
 }
 
-#' @rdname effect_shift
-#' @export
+# @rdname effect_shift
+# @export
 effect_shift_stepdown <- function(coef1, coef2) {
     assert_that(length(coef1) == length(coef2))
     n <- length(coef1)
@@ -331,8 +331,8 @@ effect_shift_stepdown <- function(coef1, coef2) {
     effect_shift_inner(n, mat, coef1, coef2, NULL)
 }
 
-#' @rdname effect_shift
-#' @export
+# @rdname effect_shift
+# @export
 effect_shift_stepup <- function(coef1, coef2) {
     effect_shift_stepdown(rev(coef2), rev(coef1))
 }
@@ -375,33 +375,33 @@ effect_shift_inner <- function(n, mat, coef1, coef2, limits=NULL) {
 effect_shift_log2 <- function(coef1, coef2)
     effect_link_log2(effect_shift(coef1, coef2))
 
-#' @rdname effect_shift
-#' @export
+# @rdname effect_shift
+# @export
 effect_shift_stepdown_log2 <- function(coef1, coef2)
     effect_link_log2(effect_shift_stepdown(coef1, coef2))
 
-#' @rdname effect_shift
-#' @export
+# @rdname effect_shift
+# @export
 effect_shift_stepup_log2 <- function(coef1, coef2)
     effect_link_log2(effect_shift_stepup(coef1, coef2))
 
 
 
-#' Goodman and Kruskall's gamma, Yule's Q
-#'
-#' Goodman and Kruskall's gamma as an effect size. Yule's Q is a special case where coef1 and coef2 both have two coefficients, and is a symmetric effect size for the interaction of two experimental factors.
-#'
-#' \code{effect_gamma_log2} is adapted to work with log2 scaled coefficients. This is almost certainly the version you want.
-#'
-#' @param coef1 Column numbers in the design matrix for the first condition, in some meaningful order.
-#'
-#' @param coef2 Corresponding column numbers for the second condition.
-#'
-#' @return
-#'
-#' An object defining how to calculate an effect size.
-#'
-#' @export
+# Goodman and Kruskall's gamma, Yule's Q
+#
+# Goodman and Kruskall's gamma as an effect size. Yule's Q is a special case where coef1 and coef2 both have two coefficients, and is a symmetric effect size for the interaction of two experimental factors.
+#
+# \code{effect_gamma_log2} is adapted to work with log2 scaled coefficients. This is almost certainly the version you want.
+#
+# @param coef1 Column numbers in the design matrix for the first condition, in some meaningful order.
+#
+# @param coef2 Corresponding column numbers for the second condition.
+#
+# @return
+#
+# An object defining how to calculate an effect size.
+#
+# @export
 effect_gamma <- function(coef1, coef2) {
     assert_that(length(coef1) == length(coef2))
     n <- length(coef1)

R/group_confects.R

@@ -20,27 +20,27 @@ group_effect_1 <- function(design, coef, effect_func, design_common=NULL) {
         get_effect = get_effect)
 }
 
-#' Group effect for differential splicing
-#'
-#' Create a group effect object to detect differential splicing (or similar). The effect size is the root sum of squared differences of the coefficient from the mean.
-#'
-#' The coefficient should represent a difference between two conditions.
-#'
-#' The idea is to detect differences in differential expression between features in a group (ie exons in a gene).
-#'
-#' @param design Design matrix.
-#'
-#' @param coef Column number in design matrix of the coefficient to be tested.
-#'
-#' @param design_common Experimental! Optional sample-level design matrix. For example, this can be used to account for a batch effect or matched samples.
-#'
-#' @return
-#'
-#' A group effect object.
-#'
-#' @seealso \code{\link{effect_rssm}}
-#'
-#' @export
+# Group effect for differential splicing
+#
+# Create a group effect object to detect differential splicing (or similar). The effect size is the root sum of squared differences of the coefficient from the mean.
+#
+# The coefficient should represent a difference between two conditions.
+#
+# The idea is to detect differences in differential expression between features in a group (ie exons in a gene).
+#
+# @param design Design matrix.
+#
+# @param coef Column number in design matrix of the coefficient to be tested.
+#
+# @param design_common Experimental! Optional sample-level design matrix. For example, this can be used to account for a batch effect or matched samples.
+#
+# @return
+#
+# A group effect object.
+#
+# @seealso \code{\link{effect_rssm}}
+#
+# @export
 group_effect_rssm <- function(design, coef, design_common=NULL) 
     group_effect_1(design, coef, effect_rssm, design_common=design_common)
 
@@ -73,7 +73,7 @@ group_effect_2 <- function(design, coef1, coef2, effect_func, design_common=NULL
 #'
 #' The coefficients should represent the expression levels in two different conditions.
 #'
-#' The _stepup and _stepdown versions may be used to look for shifts in start or end of transcription from RNA-seq or microarray data, where the observed levels are expected to be cumulative or reverse cumulative.
+# The _stepup and _stepdown versions may be used to look for shifts in start or end of transcription from RNA-seq or microarray data, where the observed levels are expected to be cumulative or reverse cumulative.
 #'
 #' @param design Design matrix.
 #'
@@ -93,13 +93,13 @@ group_effect_2 <- function(design, coef1, coef2, effect_func, design_common=NULL
 group_effect_shift <- function(design, coef1, coef2, design_common=NULL) 
     group_effect_2(design, coef1, coef2, effect_shift, design_common=design_common)
 
-#' @rdname group_effect_shift
-#' @export
+# @rdname group_effect_shift
+# @export
 group_effect_shift_stepup <- function(design, coef1, coef2, design_common=NULL) 
     group_effect_2(design, coef1, coef2, effect_shift_stepup, design_common=design_common)
 
-#' @rdname group_effect_shift
-#' @export
+# @rdname group_effect_shift
+# @export
 group_effect_shift_stepdown <- function(design, coef1, coef2, design_common=NULL) 
     group_effect_2(design, coef1, coef2, effect_shift_stepdown, design_common=design_common)
 
@@ -108,13 +108,13 @@ group_effect_shift_stepdown <- function(design, coef1, coef2, design_common=NULL
 group_effect_shift_log2 <- function(design, coef1, coef2, design_common=NULL) 
     group_effect_2(design, coef1, coef2, effect_shift_log2, design_common=design_common)
 
-#' @rdname group_effect_shift
-#' @export
+# @rdname group_effect_shift
+# @export
 group_effect_shift_stepup_log2 <- function(design, coef1, coef2, design_common=NULL) 
     group_effect_2(design, coef1, coef2, effect_shift_stepup_log2, design_common=design_common)
 
-#' @rdname group_effect_shift
-#' @export
+# @rdname group_effect_shift
+# @export
 group_effect_shift_stepdown_log2 <- function(design, coef1, coef2, design_common=NULL) 
     group_effect_2(design, coef1, coef2, effect_shift_stepdown_log2, design_common=design_common)
 
@@ -154,7 +154,7 @@ group_design_maker <- function(design, design_common=NULL) {
     memoise(group_design)
 }
 
-#' Group confects (differential exon usage, etc)
+#' Group confects (differential 5' or 3' end usage, etc)
 #'
 #' Find differential exon usage, etc.
 #'

index.md

@@ -1,6 +1,6 @@
 # | -●- Topconfects 
 
-TOP results by CONfident efFECT Size. Topconfects is an R package intended for RNA-seq or microarray Differntial Expression analysis and similar, where we are interested in placing confidence bounds on many effect sizes--one per gene--from few samples.
+TOP results by CONfident efFECT Size. Topconfects is an R package intended for RNA-seq or microarray Differntial Expression analysis and similar, where we are interested in placing confidence bounds on many effect sizes---one per gene---from few samples.
 
 Topconfects builds on [TREAT](http://bioinformatics.oxfordjournals.org/content/25/6/765.long) p-values offered by the limma and edgeR packages. It tries a range of fold changes, and uses this to rank genes by effect size while maintaining a given FDR. This also produces confidence bounds on the fold changes, with adjustment for multiple testing. See [nest_confects](reference/nest_confects.html) for details.
 
@@ -42,11 +42,13 @@ Topconfects is developed by Paul Harrison [@paulfharrison](https://twitter.com/p
 
 <br/>
 
+<!--
 ## Future work
 
 Gene-set enrichment tests. Here also the smallest p-value does not necessarily imply the greatest interest.
 
 <br/>
+-->
 
 ## References