@@ -75,30 +75,29 @@ Collate:

     'wilcox_test_class.R'

     'zzz.R'

 Depends: struct (>= 0.99.3)

-Imports: ggplot2,

- pmp,

- gridExtra,

- sp,

- scales,

- stats,

- methods,

+Imports: datasets,

+    ggplot2,

+    ggthemes,

+    grid,

+    gridExtra,

+    methods,

+    scales,

+    sp,

+    stats

+RoxygenNote: 7.0.2

+Suggests: agricolae,

+ BiocStyle,

  car,

- grid,

- reshape2,

- agricolae,

+ covr,

  emmeans,

- nlme,

- ggthemes,

+ pmp,

  ggdendro,

- datasets

-RoxygenNote: 7.0.2

-Suggests: 

- testthat,

- covr,

  knitr,

- rmarkdown,

- BiocStyle,

+ nlme,

  pls,

- Rtsne

+ reshape2,

+ rmarkdown,

+ Rtsne,

+ testthat

 VignetteBuilder: knitr

 biocViews: WorkflowStep

@@ -97,14 +97,12 @@ exportMethods(model_predict)

 exportMethods(model_reverse)

 exportMethods(model_train)

 exportMethods(run)

-import(ggdendro)

 import(ggplot2)

 import(ggthemes)

 import(grid)

 import(gridExtra)

 import(methods)

 import(pmp)

-import(reshape2)

 import(scales)

 import(stats)

 import(struct)

@@ -14,7 +14,7 @@

 #' @export HSD

 #' @examples

 #' M = HSD()

-HSD = function(alpha=0.05,mtc='fdr',formula,unblanaced=FALSE,...) {

+HSD = function(alpha=0.05,mtc='fdr',formula,unbalanced=FALSE,...) {

     out=struct::new_struct('HSD',

         alpha=alpha,

         mtc=mtc,

@@ -162,7 +162,7 @@ setMethod(f="model_apply",

             # for each combination of factors...

             out2=lapply(FF,function(x) {

-                A=HSD.test(LM,x,group = FALSE)$comparison

+                A=agricolae::HSD.test(LM,x,group = FALSE)$comparison

                 if (ALIAS) {

                     A[!is.na(A)]=NA # replace with NA if alias are present

                 }

@@ -9,7 +9,7 @@

 #' @export PCA

 #' @examples

 #' M = PCA()

-PCA = function(umber_components=2,...) {

+PCA = function(number_components=2,...) {

     out=struct::new_struct('PCA',

         number_components=number_components,

         ...)

@@ -79,7 +79,7 @@ setMethod(f="chart_plot",

 #' @param label_filter Only include labels for samples in the group specified by label_filter.

 #' If zero length then all labels will be included.

 #' @param label_factor The sample_meta column to use for labelling the samples.

-#' If zero length then the rownames will be used.

+#' If 'rownames' then the rownames will be used.

 #' @param label_size The text size of the labels.NB ggplot units, not font size units.

 #' Default 3.88.

 #' @param ... additional slots and values passed to struct_class

@@ -94,10 +94,10 @@ pca_scores_plot = function(

     factor_name,

     ellipse='all',

     label_filter=character(0),

-    label_factor=character(0),

+    label_factor='rownames',

     label_size=3.88,

     ...) {

-    out=struct::new_struct(pca_scores_plot,

+    out=struct::new_struct('pca_scores_plot',

         components=components,

         points_to_label=points_to_label,

         factor_name=factor_name,

@@ -356,6 +356,8 @@ pca_biplot_plot = function(

     prototype = list(name='Feature boxplot',

         description='plots a boxplot of a chosen feature for each group of a DatasetExperiment.',

         type="boxlot",

+        .params=c('components','points_to_label','factor_name','scale_factor','style','label_features'),

+

         components=entity(name='Components to plot',

             value=c(1,2),

             type='numeric',

@@ -483,7 +485,7 @@ setMethod(f="chart_plot",

 #' @include PCA_class.R

 #' @examples

 #' C = pca_loadings_plot()

-pca_loadings_plot = function(components=c(1,2),style='points',label_featurs=FALSE,...) {

+pca_loadings_plot = function(components=c(1,2),style='points',label_features=FALSE,...) {

     out=struct::new_struct('pca_loadings_plot',

         components=components,

         style=style,

@@ -629,17 +631,23 @@ setMethod(f="chart_plot",

 #' pca_dstat_plot class

 #'

-#' Line plot showing percent variance and cumulative percent variance for the computed components.

+#' Bar chart showing mahalanobis distance from the mean in PCA scores space. A threshold is

+#' plotted at a chosen confidence as an indicator for rejecting outliers.

 #'

 #' @import struct

+#' @param number_components The number of components to use.

+#' @param alpha The confidence level to plot.

 #' @param ... additional slots and values passed to struct_class

 #' @return struct object

 #' @export PCA_dstat

 #' @include PCA_class.R

 #' @examples

 #' C = PCA_dstat()

-PCA_dstat = function(...) {

-    out=struct::new_struct('PCA_dstat',...)

+PCA_dstat = function(number_components=2,alpha=0.05,...) {

+    out=struct::new_struct('PCA_dstat',

+        number_components=number_components,

+        alpha=alpha,

+        ...)

     return(out)

 }

@@ -652,6 +660,8 @@ PCA_dstat = function(...) {

     prototype = list(name='d-statistic plot',

         description='a bar chart of the d-statistics for samples in the input PCA model',

         type="bar",

+        .params=c('number_components','alpha'),

+

         number_components=entity(value = 2,

             name = 'number of principal components',

             description = 'number of principal components to use for the plot',

@@ -39,7 +39,7 @@ plsda_scores_plot = function(components=c(1,2),points_to_label='none',factor_nam

         description='scatter plot of PLSDA component scores',

         type="scatter",

         libraries=c('pls','ggplot2'),

-        .params=c('components','points_to_label,factor_name'),

+        .params=c('components','points_to_label','factor_name','groups'),

         components=entity(name='Components to plot',

             value=c(1,2),

@@ -8,7 +8,7 @@

 #' @export PLSR

 #' @examples

 #' M = PLSR()

-PLSR = function(number_components=c(1,2),factor_name,...) {

+PLSR = function(number_components=2,factor_name,...) {

     out=struct::new_struct('PLSR',

         number_components=number_components,

         factor_name=factor_name,

@@ -16,7 +16,7 @@

 #' @return struct object

 #' @export balanced_accuracy

 balanced_accuracy = function(...) {

-    out=struct::new_struct(out,...)

+    out=struct::new_struct('balanced_accuracy',...)

     return(out)

 }

@@ -48,7 +48,7 @@ blank_filter = function(fold_change=20,blank_label='blank',qc_label='QC',factor_

         type = 'filter',

         predicted = 'filtered',

         libraries='pmp',

-        .params=c('blank_label','qc_label','factor_name','fraction_in_blank'),

+        .params=c('fold_change','blank_label','qc_label','factor_name','fraction_in_blank'),

         .outputs=c('filtered','flags'),

         blank_label=ents$blank_label,

@@ -25,7 +25,7 @@

 classical_lsq = function(alpha=0.05,mtc='fdr',factor_names,intercept=TRUE,...) {

-    out=struct::new_struct(out,

+    out=struct::new_struct('classical_lsq',

         alpha = alpha,

         mtc = mtc,

         factor_names = factor_names,

@@ -57,11 +57,11 @@ classical_lsq = function(alpha=0.05,mtc='fdr',factor_names,intercept=TRUE,...) {

         type="univariate",

         predicted='p_value',

         .params=c('alpha','mtc','factor_names','intercept'),

-        .outputs=c('coefficients','p_value','significant','r_squared','adj_r-squared'),

+        .outputs=c('coefficients','p_value','significant','r_squared','adj_r_squared'),

         intercept=entity(name='Include intercept',

             type='logical',

-            description='TRUE or FALSE to include the intercept term when fitting the model_',

+            description='TRUE or FALSE to include the intercept term when fitting the model',

             value=TRUE

         ),

@@ -28,8 +28,14 @@

 #' @param ... additional slots and values passed to struct_class

 #' @return struct object

 #' @export confounders_clsq

-confounders_clsq = function(alpha=0.05,mtc='fdr',factor_name,confounding_factors,threshold,...) {

-    out=struct::new_struct(out,...)

+confounders_clsq = function(alpha=0.05,mtc='fdr',factor_name,confounding_factors,threshold=0.15,...) {

+    out=struct::new_struct('confounders_clsq',

+        alpha=alpha,

+        mtc=mtc,

+        factor_name=factor_name,

+        confounding_factors=confounding_factors,

+        threshold=threshold,

+        ...)

     return(out)

 }

@@ -57,7 +63,7 @@ confounders_clsq = function(alpha=0.05,mtc='fdr',factor_name,confounding_factors

         type="univariate",

         predicted='p_value',

         .params=c('alpha','mtc','factor_name','confounding_factors','threshold'),

-        .outputs=c('coefficients','p_value','significant','percent_change','significant'),

+        .outputs=c('coefficients','p_value','significant','percent_change','potential_confounders'),

         threshold=entity(name='Confounding factor threshold',

             type='numeric',

@@ -97,12 +103,12 @@ setMethod(f="model_apply",

     definition=function(M,D)

     {

         # classical least squares model

-        clsq=classical_lsq(intercept=TRUE,alpha=M$alpha,mtc=M$mtc)

+        clsq=classical_lsq(intercept=TRUE,alpha=M$alpha,mtc=M$mtc,factor_names='dummy')

         # make list of all factors

         factor_names=c(M$factor_name,M$confounding_factors)

-        # do a regression including the main factor and the counfounders one at a time

+        # do a regression including the main factor and the confounders one at a time

         temp=matrix(NA,nrow=ncol(D$data),ncol=length(factor_names)) # coefficients

         pvals=temp # p-values

         nm=character(length(factor_names))

@@ -110,12 +116,12 @@ setMethod(f="model_apply",

             fn=unique(c(factor_names[1],factor_names[i]))

             # for each factor name check the na count

-            FF=filter_na_count(threshold=2)

+            FF=filter_na_count(threshold=2,factor_name='dummy')

             excl=matrix(NA,nrow=ncol(D$data),ncol=length(fn))

             colnames(excl)=fn

             for (k in fn) {

                 if (is.factor(D$sample_meta[,k])) {

-                    FF$factor_name=k

+                    FF$factor_name=k # replace dummy factor name

                     FF=model_apply(FF,D)

                     excl[,k]=FF$flags$flags

                 } else {

@@ -131,7 +137,7 @@ setMethod(f="model_apply",

                 excl=fn #

             }

-            clsq$factor_names=excl

+            clsq$factor_names=excl # put real factor names instead of dummy

             clsq=model_apply(clsq,D)

             nm[i]=paste0(fn,collapse='_')

@@ -195,7 +201,7 @@ setMethod(f="model_apply",

 #' @param ... additional slots and values passed to struct_class

 #' @return struct object

 #' @export confounders_lsq.barchart

-confounders_lsq.barchart = function(feature_to_plot,threshold,...) {

+confounders_lsq.barchart = function(feature_to_plot,threshold=10,...) {

     out=struct::new_struct('confounders_lsq.barchart',

         feature_to_plot=feature_to_plot,

         threshold=threshold,

@@ -284,7 +290,7 @@ setMethod(f="chart_plot",

 #' @param ... additional slots and values passed to struct_class

 #' @return struct object

 #' @export confounders_lsq.boxplot

-confounders_lsq.boxplot = function(threshold,...) {

+confounders_lsq.boxplot = function(threshold=10,...) {

     out=struct::new_struct('confounders_lsq.boxplot',

         threshold=threshold,

         ...)

@@ -17,7 +17,7 @@

 #' @return A struct chart object

 #' @export feature_boxplot

 feature_boxplot = function(label_outliers=TRUE,feature_to_plot,factor_name,show_counts=TRUE,...) {

-    out=struct::new_struct('feature_box_plot',

+    out=struct::new_struct('feature_boxplot',

         label_outliers=label_outliers,

         feature_to_plot=feature_to_plot,

         factor_name=factor_name,

@@ -248,6 +248,10 @@ setMethod(f="chart_plot",

 #' chart_plot(C,D)

 #'

 #' @import struct

+#' @param label_outliers TRUE or FALSE to label outliers on the plot.

+#' @param by_sample TRUE to plot missing values by sample, or FALSE to plot for features.

+#' @param factor_name The sample_meta column to use.

+#' @param show_counts TRUE to show a count of the number of items used to create the boxplot on the chart.

 #' @param ... additional slots and values passed to struct_class

 #' @return struct object

 #' @export mv_boxplot

@@ -275,6 +279,8 @@ mv_boxplot = function(label_outliers=TRUE,by_sample=TRUE,factor_name,show_counts

     prototype = list(name='Missing value boxplots',

         description='Histogram ofmissing values per sample/feature.',

         type="histogram",

+        .params=c('label_outliers','by_sample','factor_name','show_counts'),

+

         label_outliers=entity(name='Label outliers',

             value=TRUE,

             type='logical',

@@ -501,10 +507,10 @@ setMethod(f="chart_plot",

 #' @param ... additional slots and values passed to struct_class

 #' @return struct object

 #' @export DatasetExperiment.boxplot

-DatasetExperiment.boxplot = function(factor_name,by_sample=TRUE,per_class=TRUE,number,...) {

+DatasetExperiment.boxplot = function(factor_name,by_sample=TRUE,per_class=TRUE,number=50,...) {

     out=struct::new_struct('DatasetExperiment.boxplot',

         factor_name=factor_name,

-        by_sample-by_sample,

+        by_sample=by_sample,

         per_class=per_class,

         number=number,

         ...)

@@ -633,6 +639,7 @@ compare_dist = function(factor_name,...) {

     prototype = list(name='Compare distributions',

         description='Distributions and box plots to compare two datasets',

         type="mixed",

+        .params=c('factor_name'),

         factor_name=entity(name='Factor name',

             value='factor',

             type='character',

@@ -725,16 +732,15 @@ setMethod(f="chart_plot",

 #'

 #' plots a DatasetExperiment as a heatmap

 #'

-#' @import struct

-#' @import reshape2

 #' @param ... additional slots and values passed to struct_class

+#' @param na_colour A hex colour code to use for missing values

 #' @return struct object

 #' @export DatasetExperiment.heatmap

 #' @examples

 #' C = DatasetExperiment.heatmap()

-DatasetExperiment.heatmap = function(...) {

-    out=.DatasetExperiment.heatmap()

-    out=struct::new_struct(out,...)

+DatasetExperiment.heatmap = function(na_colour='#FF00E4',...) {

+    out=struct::new_struct('DatasetExperiment.heatmap',

+        na_colour=na_colour,...)

     return(out)

 }

@@ -749,6 +755,8 @@ DatasetExperiment.heatmap = function(...) {

     prototype = list(name='DatasetExperiment heatmap',

         description='plots a heatmap of a DatasetExperiment',

         type="scatter",

+        libraries='reshape2',

+        .params=c('na_colour'),

         na_colour=entity(name='NA colour',

             value='#FF00E4',

@@ -758,14 +766,13 @@ DatasetExperiment.heatmap = function(...) {

     )

 )

-#' @param ... additional slots and values passed to struct_class

 #' @export

 #' @template chart_plot

 setMethod(f="chart_plot",

     signature=c("DatasetExperiment.heatmap",'DatasetExperiment'),

     definition=function(obj,dobj)

     {

-        X=melt(as.matrix(dobj$data))

+        X=reshape2::melt(as.matrix(dobj$data))

         colnames(X)=c('Sample','Feature','Peak area')

         p=ggplot(data=X,aes(x=`Feature`,y=`Sample`,fill=`Peak area`)) + geom_raster() +

             scale_colour_Publication()+

@@ -19,7 +19,7 @@

 #' @return struct object

 #' @export filter_by_name

 filter_by_name = function(mode='exclude',dimension='sample',names,...) {

-    out=struct::new_struct(filter_by_name,

+    out=struct::new_struct('filter_by_name',

         mode=mode,

         dimension=dimension,

         names=names,

@@ -16,7 +16,7 @@

 #' @return struct object

 #' @export filter_smeta

 filter_smeta = function(mode='include',levels,factor_name,...) {

-    out=struct::new_struct(filter_smeta,

+    out=struct::new_struct('filter_smeta',

         mode=mode,

         levels=levels,

         factor_name=factor_name,

@@ -75,7 +75,7 @@ setMethod(f="model_apply",

         } else {

             stop('mode must be "include" or "exclude"')

         }

-        D=D[!out,,drop=FALSE]

+        D=D[!out,]

         # drop excluded levels from factors

         D$sample_meta=droplevels(D$sample_meta)

         output_value(M,'filtered')=D

@@ -99,8 +99,8 @@ setMethod(f="model_apply",

         s=p<M$alpha

         names(s)=colnames(X)

-        M$p_value=p

-        M$significant=s

+        M$p_value=as.data.frame(p)

+        M$significant=as.data.frame(s)

         return(M)

     }

@@ -8,7 +8,7 @@

 #' @examples

 #' M = glog_transform()

 glog_transform = function(qc_label='QC',factor_name,...) {

-    out=struct::new_struct('glog',

+    out=struct::new_struct('glog_transform',

         qc_label=qc_label,

         factor_name=factor_name,

         ...)

@@ -27,8 +27,8 @@ glog_transform = function(qc_label='QC',factor_name,...) {

         lambda_opt='numeric'

     ),

-    prototype=list(name = 'generalised logarithm transform',

-        description = 'applies a glog tranform using using QC samples as reference samples.',

+    prototype=list(name = 'Generalised logarithm transform',

+        description = 'Applies a glog transform using using QC samples as reference samples.',

         type = 'normalisation',

         predicted = 'transformed',

         libraries = 'pmp',

@@ -101,7 +101,6 @@ setMethod(f="model_apply",

 #' @param ... additional slots and values passed to struct_class

 #' @return struct object

 #' @export hca_dendrogram

-#' @import ggdendro

 #' @include hca_class.R

 #' @examples

 #' C = hca_dendrogram()

@@ -113,7 +112,8 @@ hca_dendrogram = function(...) {

 .hca_dendrogram<-setClass(

     "hca_dendrogram",

-    contains='chart'

+    contains='chart',

+    prototype = list(libraries='ggdendro')

 )

 #' @export

@@ -2,14 +2,14 @@

 #'

 #' Applies a k-nearest neighbour approach to impute missing values.

 #' @param neighbours The number of neighbours to use for imputation.

-#' @param sample_max Maximum proportion of missing values in any sample.

-#' @param feature_max Maximum proportion of missing values in any feature.

+#' @param sample_max Maximum percentage of missing values in any sample. Default = 50.

+#' @param feature_max Maximum percentage of missing values in any feature. Default = 50.

 #' @param ... additional slots and values passed to struct_class

 #' @return struct object

 #' @export knn_impute

 #' @examples

 #' M = knn_impute()

-knn_impute = function(neighbours=5,sample_max=0.5,feature_max=0.5,...) {

+knn_impute = function(neighbours=5,sample_max=50,feature_max=50,...) {

     out=struct::new_struct('knn_impute',

         neighbours=neighbours,

         sample_max=sample_max,

@@ -71,7 +71,7 @@ setMethod(f="model_apply",

         smeta=D$sample_meta

         x=D$data

-        imputed = mv_imputation(t(as.matrix(x)),method='knn',k = opt$neighbours,rowmax=opt$feature_max/100,colmax=opt$sample_max/100,maxp = NULL,FALSE)

+        imputed = pmp::mv_imputation(t(as.matrix(x)),method='knn',k = opt$neighbours,rowmax=opt$feature_max/100,colmax=opt$sample_max/100,maxp = NULL,FALSE)

         D$data = as.data.frame(t(imputed))

         output_value(M,'imputed') = D

@@ -12,10 +12,10 @@

 #' @export linear_model

 #' @examples

 #' M = linear_model()

-linear_model = function(formula,na_action='na_omit',contrasts=list(),...) {

+linear_model = function(formula,na_action='na.omit',contrasts=list(),...) {

     out=struct::new_struct('linear_model',

         formula=formula,

-        na_action=nna_action,

+        na_action=na_action,

         contrasts=contrasts,

         ...)

     return(out)

@@ -67,7 +67,7 @@ setMethod(f="model_apply",

             var_names_ex=var_names

         }

-        FF=full_fact(var_names_ex)

+        FF=structToolbox:::full_fact(var_names_ex)

         FF=apply(FF,1,function(x) var_names_ex[x==1])

         FF=FF[-1]

@@ -78,7 +78,7 @@ setMethod(f="model_apply",

             dona=FALSE

             testlm=tryCatch({ # if any warnings/messages set p-values to NA as unreliable

-                LM=lme(lmer_formula$f,random=lmer_formula$random,method='ML',data=temp,na.action=na.omit)

+                LM=nlme::lme(lmer_formula$f,random=lmer_formula$random,method='ML',data=temp,na.action=na.omit)

             }, warning=function(w) {

                 NA

             }, message=function(m) {

@@ -17,7 +17,7 @@ mv_feature_filter = function(threshold=20,qc_label='QC',method='QC',factor_name,

         threshold=threshold,

         qc_label=qc_label,

         method=method,

-        factor_name,

+        factor_name=factor_name,

         ...)

     return(out)

 }

@@ -87,10 +87,10 @@ setMethod(f="model_train",

         s=strsplit(opt$method,'_')[[1]][1]

-        filtered = filter_peaks_by_fraction(t(x), min_frac = opt$threshold/100, classes=smeta[[M$factor_name]], method=s,qc_label=opt$qc_label)

+        filtered = pmp::filter_peaks_by_fraction(t(x), min_frac = opt$threshold/100, classes=smeta[[M$factor_name]], method=s,qc_label=opt$qc_label,remove_peaks = FALSE)

         #D$data = as.data.frame(t(filtered$df))

-        flags<-data.frame(filtered$flags)

+        flags<-data.frame(attributes(filtered)$flags)

         output_value(M,'flags') = flags

@@ -58,11 +58,11 @@ setMethod(f="model_apply",

         smeta=D$sample_meta

         x=D$data

-        filtered = filter_samples_by_mv(x,max_perc_mv=opt$mv_threshold/100,D$sample_meta[,1])

+        filtered = pmp::filter_samples_by_mv(x,max_perc_mv=opt$mv_threshold/100,D$sample_meta[,1],remove_samples = FALSE)

-        flags<-data.frame(filtered$flags)

+        flags<-data.frame(attributes(filtered)$flags)

-        D=D[flags$flags==1,,drop=FALSE]

+        D=D[flags$filter_samples_by_mv_flags==1,,drop=FALSE]

         output_value(M,'filtered') = D

         output_value(M,'flags') = flags

@@ -1,4 +1,4 @@

-#' Probabilistic Quotient Nomalisation

+#' Probabilistic Quotient Normalisation

 #'

 #' Applies PQN using QC samples as reference samples

 #' @param qc_label = The label for qc samples in the chosen sample_meta column.

@@ -8,8 +8,11 @@

 #' @export pqn_norm

 #' @examples

 #' M = pqn_norm()

-pqn_norm = function(...) {

-    out=struct::new_struct('pqn_norm',qc_label='QC',factor_name,...)

+pqn_norm = function(qc_label='QC',factor_name=factor_name,...) {

+    out=struct::new_struct('pqn_norm',

+        qc_label=qc_label,

+        factor_name=factor_name,

+        ...)

     return(out)

 }

@@ -59,11 +62,11 @@ setMethod(f="model_apply",

         smeta=D$sample_meta

         x=D$data

-        normalised = pqn_normalisation(t(x), classes=smeta[,M$factor_name],qc_label=opt$qc_label) # operates on transpose of x

-        D$data = as.data.frame(t(normalised$df))

+        normalised = pmp::pqn_normalisation(t(x), classes=smeta[,M$factor_name],qc_label=opt$qc_label) # operates on transpose of x

+        D$data = as.data.frame(t(normalised))

         output_value(M,'normalised') = D

-        output_value(M,'coeff') = data.frame('coeff'=normalised$coef,row.names = rownames(x))

+        output_value(M,'coeff') = data.frame('coeff'=attributes(normalised)$flags,row.names = rownames(x))

         return(M)

     }

@@ -12,7 +12,11 @@

 #' M = rsd_filter()

 #'

 rsd_filter = function(rsd_threshold=20,qc_label='QC',factor_name,...) {

-    out=struct::new_struct('rsd_filter',...)

+    out=struct::new_struct('rsd_filter',

+        rsd_threshold=rsd_threshold,

+        qc_label=qc_label,

+        factor_name=factor_name,

+        ...)

     return(out)

 }

@@ -23,14 +27,15 @@ rsd_filter = function(rsd_threshold=20,qc_label='QC',factor_name,...) {

         qc_label='entity',

         factor_name='entity',

         filtered='entity',

-        flags='entity'

+        flags='entity',

+        rsd_qc='entity'

     ),

     prototype=list(name = 'RSD filter',

         description = 'Filters features by calculating the relative standard deviation (RSD) for the QC samples and removing features with RSD greater than the threshold.',

         type = 'filter',

         predicted = 'filtered',

         .params=c('rsd_threshold','qc_label','factor_name'),

-        .outputs=c('filtered','flags'),

+        .outputs=c('filtered','flags','rsd_qc'),

         rsd_threshold=entity(name = 'RSD threhsold',

             description = 'Features with RSD greater than the threshold are removed.',

@@ -56,6 +61,11 @@ rsd_filter = function(rsd_threshold=20,qc_label='QC',factor_name,...) {

             description = 'RSD and a flag indicating whether the feature was rejected by the filter or not.',

             type='data.frame',

             value=data.frame()

+        ),

+        rsd_qc=entity(name = 'RSD',

+            description = 'The calculated RSD of the QC class',

+            type='data.frame',

+            value=data.frame()

         )

     )

 )

@@ -69,13 +79,14 @@ setMethod(f="model_apply",

         opt=param_list(M)

         smeta=D$sample_meta

         x=D$data

-        rsd_filtered = filter_peaks_by_rsd(t(x), max_rsd = opt$rsd_threshold, classes=smeta[[opt$factor_name]], qc_label=opt$qc_label)

+        rsd_filtered = pmp::filter_peaks_by_rsd(t(x), max_rsd = opt$rsd_threshold, classes=smeta[[opt$factor_name]], qc_label=opt$qc_label,remove_peaks=FALSE)

-        flags<-data.frame(rsd_filtered$flags)

-        D=D[,flags[,2]==1,drop=FALSE]

+        flags<-attributes(rsd_filtered)$flags

+        D=D[,flags[,2]==1]

         output_value(M,'filtered') = D

-        output_value(M,'flags') = data.frame(rsd_filtered$flags,stringsAsFactors = F)

+        output_value(M,'flags') = data.frame('rsd_flags'=flags[,2])

+        output_value(M,'rsd_qc') = data.frame('rsd_qc'=flags[,1])

         return(M)

     }

 )

@@ -115,10 +126,10 @@ setMethod(f="chart_plot",

     {

         t=param_value(dobj,'rsd_threshold')

         A=output_value(dobj,'flags')

-        A$rsd_QC=log2(A$rsd_QC)

+        A$rsd_qc=log2(dobj$rsd_qc[,1])

         A$features=factor(A$rsd_flags,levels=c(1,0),labels=c('accepted','rejected'))

-        out=ggplot(data=A, aes_(x=~rsd_QC,fill=~features)) +

+        out=ggplot(data=A, aes_(x=~rsd_qc,fill=~features)) +

             geom_histogram(boundary=log2(t),color='white') +

             xlab('log2(RSD), QC samples') +

             ylab('Count') +

@@ -49,7 +49,7 @@ ttest = function(alpha=0.05,mtc='fdr',factor_names,paired=FALSE,paired_factor=ch

         type="univariate",

         predicted='p_value',

         stato_id="STATO:0000304",

-        .params=c('alpha','mtc','factor_name','paired','paired_factor'),

+        .params=c('alpha','mtc','factor_names','paired','paired_factor'),

         .outputs=c('t_statistic','p_value','dof','significant','conf_int','estimates'),

         factor_names=entity(name='Factor names',

@@ -8,7 +8,7 @@ DatasetExperiment.boxplot(

   factor_name,

   by_sample = TRUE,

   per_class = TRUE,

-  number,

+  number = 50,

   ...

 )

 }

@@ -4,9 +4,11 @@

 \alias{DatasetExperiment.heatmap}

 \title{DatasetExperiment.heatmap class}

 \usage{

-DatasetExperiment.heatmap(...)

+DatasetExperiment.heatmap(na_colour = "#FF00E4", ...)

 }

 \arguments{

+\item{na_colour}{A hex colour code to use for missing values}

+

 \item{...}{additional slots and values passed to struct_class}

 }

 \value{

@@ -4,7 +4,7 @@

 \alias{HSD}

 \title{HSD model class}

 \usage{

-HSD(alpha = 0.05, mtc = "fdr", formula, unblanaced = FALSE, ...)

+HSD(alpha = 0.05, mtc = "fdr", formula, unbalanced = FALSE, ...)

 }

 \arguments{

 \item{alpha}{The p-value threshold. Default alpha = 0.05.}

@@ -13,9 +13,9 @@ HSD(alpha = 0.05, mtc = "fdr", formula, unblanaced = FALSE, ...)

 \item{formula}{The formula to use. See \code{lm} for details.}

-\item{...}{additional slots and values passed to struct_class}

-

 \item{unbalanced}{TRUE or FALSE to apply correction for unbalanced designs. Default is FALSE.}

+

+\item{...}{additional slots and values passed to struct_class}

 }

 \value{

 struct object

@@ -4,12 +4,12 @@

 \alias{PCA}

 \title{PCA model class}

 \usage{

-PCA(umber_components = 2, ...)

+PCA(number_components = 2, ...)

 }

 \arguments{

-\item{...}{additional slots and values passed to struct_class}

-

 \item{number_components}{The number of principal components to retain}

+

+\item{...}{additional slots and values passed to struct_class}

 }

 \value{

 struct object

@@ -4,16 +4,21 @@

 \alias{PCA_dstat}

 \title{pca_dstat_plot class}

 \usage{

-PCA_dstat(...)

+PCA_dstat(number_components = 2, alpha = 0.05, ...)

 }

 \arguments{

+\item{number_components}{The number of components to use.}

+

+\item{alpha}{The confidence level to plot.}

+

 \item{...}{additional slots and values passed to struct_class}

 }

 \value{

 struct object

 }

 \description{

-Line plot showing percent variance and cumulative percent variance for the computed components.

+Bar chart showing mahalanobis distance from the mean in PCA scores space. A threshold is

+plotted at a chosen confidence as an indicator for rejecting outliers.

 }

 \examples{

 C = PCA_dstat()

@@ -4,7 +4,7 @@

 \alias{PLSR}

 \title{PLSR model class}

 \usage{

-PLSR(number_components = c(1, 2), factor_name, ...)

+PLSR(number_components = 2, factor_name, ...)

 }

 \arguments{

 \item{number_components}{The number of PLS components to calculate.}

@@ -9,7 +9,7 @@ confounders_clsq(

   mtc = "fdr",

   factor_name,

   confounding_factors,

-  threshold,

+  threshold = 0.15,

   ...

 )

 }

@@ -4,7 +4,7 @@

 \alias{confounders_lsq.barchart}

 \title{barchart of percent change}

 \usage{

-confounders_lsq.barchart(feature_to_plot, threshold, ...)

+confounders_lsq.barchart(feature_to_plot, threshold = 10, ...)

 }

 \arguments{

 \item{feature_to_plot}{the name or index of the feature to be plotted}

@@ -4,7 +4,7 @@

 \alias{confounders_lsq.boxplot}

 \title{boxplot of percent change}

 \usage{

-confounders_lsq.boxplot(threshold, ...)

+confounders_lsq.boxplot(threshold = 10, ...)

 }

 \arguments{

 \item{threshold}{the threshold to be plotted (in \%)}

@@ -4,14 +4,14 @@

 \alias{knn_impute}

 \title{knn missing value imputation}

 \usage{

-knn_impute(neighbours = 5, sample_max = 0.5, feature_max = 0.5, ...)

+knn_impute(neighbours = 5, sample_max = 50, feature_max = 50, ...)

 }

 \arguments{

 \item{neighbours}{The number of neighbours to use for imputation.}

-\item{sample_max}{Maximum proportion of missing values in any sample.}

+\item{sample_max}{Maximum percentage of missing values in any sample. Default = 50.}

-\item{feature_max}{Maximum proportion of missing values in any feature.}

+\item{feature_max}{Maximum percentage of missing values in any feature. Default = 50.}

 \item{...}{additional slots and values passed to struct_class}

 }

@@ -4,7 +4,7 @@

 \alias{linear_model}

 \title{linear model class}

 \usage{

-linear_model(formula, na_action = "na_omit", contrasts = list(), ...)

+linear_model(formula, na_action = "na.omit", contrasts = list(), ...)

 }

 \arguments{

 \item{formula}{The formula to use.}

@@ -13,14 +13,13 @@ mv_boxplot(

 )

 }

 \arguments{