Package 'structToolbox'

Description

Analysis of Variance (ANOVA) is a univariate method used to analyse the difference among group means. Multiple test corrected p-values are computed to indicate significance for each feature.

Usage

ANOVA(alpha = 0.05, mtc = "fdr", formula, ss_type = "III", ...)

Arguments

Details

This object makes use of functionality from the following packages:

• car

Value

A ANOVA object with the following output slots:

Inheritance

A ANOVA object inherits the following struct classes:

⁠[ANOVA]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Fox J, Weisberg S (2019). An R Companion to Applied Regression, Third edition. Sage, Thousand Oaks CA. https://socialsciences.mcmaster.ca/jfox/Books/Companion/.

Examples

M = ANOVA(
      alpha = 0.05,
      mtc = "fdr",
      formula = y ~ x,
      ss_type = "III")

D = iris_DatasetExperiment()
M = ANOVA(formula=y~Species)
M = model_apply(M,D)

Description

Convert the outputs of the input model into a data.frame.

Usage

## S4 method for signature 'filter_na_count'
as_data_frame(M)

## S4 method for signature 'ttest'
as_data_frame(M)

## S4 method for signature 'wilcox_test'
as_data_frame(M)

Arguments

Value

A data.frame of model outputs

Examples

D = iris_DatasetExperiment()
M = filter_na_count(threshold=50,factor_name='Species')
M= model_apply(M,D)
df = as_data_frame(M)

Description

The area under the ROC curve of a classifier is estimated using the trapezoid method.

Usage

AUC(...)

Arguments

Value

A AUC object. This object has no output slots.

Inheritance

A AUC object inherits the following struct classes:

⁠[AUC]⁠ >> ⁠[metric]⁠ >> ⁠[struct_class]⁠

Examples

M = AUC()

D = iris_DatasetExperiment()
XCV = kfold_xval(folds=5,factor_name='Species') *
      (mean_centre() + PLSDA(number_components=2,factor_name='Species'))
MET = AUC()
XCV = run(XCV,D,MET)

Description

Each variable/feature is mean centred and scaled by the standard deviation. The transformed variables have zero-mean and unit-variance.

Usage

autoscale(mode = "data", ...)

Arguments

Value

A autoscale object with the following output slots:

Inheritance

A autoscale object inherits the following struct classes:

⁠[autoscale]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = autoscale(
      mode = "data")

D = iris_DatasetExperiment()
M = autoscale()
M = model_train(M,D)
M = model_predict(M,D)

Description

Balanced Accuracy is the average proportion of correctly classified samples across all groups.

Usage

balanced_accuracy(...)

Arguments

Value

A balanced_accuracy object. This object has no output slots.

Inheritance

A balanced_accuracy object inherits the following struct classes:

⁠[balanced_accuracy]⁠ >> ⁠[metric]⁠ >> ⁠[struct_class]⁠

Examples

M = balanced_accuracy()

D = iris_DatasetExperiment()
XCV = kfold_xval(folds=5,factor_name='Species') *
      (mean_centre() + PLSDA(number_components=2,factor_name='Species'))
MET = balanced_accuracy()
XCV = run(XCV,D,MET)

Description

A blank filter filters features by comparing the median intensity of blank samples to the median intensity of samples. Features where the relative intensity (fold change) is not large when compared to the blank are removed. The number of times a feature is detected across all blank samples may also be considered. If the feature is not detected in a high enough proportion of the blanks then it is not removed.

Usage

blank_filter(
  fold_change = 20,
  blank_label = "blank",
  qc_label = "QC",
  factor_name,
  fraction_in_blank = 0,
  ...
)

Arguments

Details

This object makes use of functionality from the following packages:

• pmp

Value

A blank_filter object with the following output slots:

Inheritance

A blank_filter object inherits the following struct classes:

⁠[blank_filter]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Jankevics A, Lloyd GR, Weber RJM (????). pmp: Peak Matrix Processing and signal batch correction for metabolomics datasets. R package version 1.15.1.

Examples

M = blank_filter(
      fold_change = 20,
      blank_label = "Blank",
      qc_label = "QC",
      factor_name = "V1",
      fraction_in_blank = 0)

D = iris_DatasetExperiment()
M = blank_filter(fold_change=2,
                 factor_name='Species',
                 blank_label='setosa',
                 qc_label='versicolor')
M = model_apply(M,D)

Description

A histogram of the calculated fold changes for the blank filter (median samples divided by median blanks)

Usage

blank_filter_hist(...)

Arguments

Value

A blank_filter_hist object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A blank_filter_hist object inherits the following struct classes:

⁠[blank_filter_hist]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = blank_filter_hist()

C = blank_filter_hist()

Description

In bootstrap resampling a subset of samples is selected at random with replacement to form a training set. Any sample not selected for training is included in the test set. This process is repeated many times, and performance metrics are computed for each repetition.

Usage

bootstrap(number_of_repetitions = 100, collect, ...)

Arguments

Value

A bootstrap object with the following output slots:

Inheritance

A bootstrap object inherits the following struct classes:

⁠[bootstrap]⁠ >> ⁠[resampler]⁠ >> ⁠[iterator]⁠ >> ⁠[struct_class]⁠

Examples

M = bootstrap(
      number_of_repetitions = 10,
      collect = "vip")

I = bootstrap(number_of_repetitions = 10, collect = 'vip')

Description

Calculate metric

Usage

## S4 method for signature 'AUC'
calculate(obj, Y, Yhat)

## S4 method for signature 'balanced_accuracy'
calculate(obj, Y, Yhat)

## S4 method for signature 'r_squared'
calculate(obj, Y, Yhat)

Arguments

Value

a modified metric object

Examples

MET = metric()
calculate(MET)

Description

Plots a chart object

Usage

## S4 method for signature 'dfa_scores_plot,DFA'
chart_plot(obj, dobj)

## S4 method for signature 'scatter_chart,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'pca_correlation_plot,PCA'
chart_plot(obj, dobj)

## S4 method for signature 'pca_scores_plot,PCA'
chart_plot(obj, dobj)

## S4 method for signature 'pca_biplot,PCA'
chart_plot(obj, dobj)

## S4 method for signature 'pca_loadings_plot,PCA'
chart_plot(obj, dobj)

## S4 method for signature 'pca_scree_plot,PCA'
chart_plot(obj, dobj)

## S4 method for signature 'pca_dstat_plot,PCA'
chart_plot(obj, dobj)

## S4 method for signature 'plsr_prediction_plot,PLSR'
chart_plot(obj, dobj)

## S4 method for signature 'plsr_residual_hist,PLSR'
chart_plot(obj, dobj)

## S4 method for signature 'plsr_qq_plot,PLSR'
chart_plot(obj, dobj)

## S4 method for signature 'plsr_cook_dist,PLSR'
chart_plot(obj, dobj)

## S4 method for signature 'pls_scores_plot,PLSR'
chart_plot(obj, dobj)

## S4 method for signature 'plsda_predicted_plot,PLSDA'
chart_plot(obj, dobj)

## S4 method for signature 'plsda_roc_plot,PLSDA'
chart_plot(obj, dobj)

## S4 method for signature 'pls_vip_plot,PLSR'
chart_plot(obj, dobj)

## S4 method for signature 'pls_regcoeff_plot,PLSR'
chart_plot(obj, dobj)

## S4 method for signature 'blank_filter_hist,blank_filter'
chart_plot(obj, dobj)

## S4 method for signature 'confounders_lsq_barchart,confounders_clsq'
chart_plot(obj, dobj)

## S4 method for signature 'confounders_lsq_boxplot,confounders_clsq'
chart_plot(obj, dobj)

## S4 method for signature 'feature_boxplot,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'mv_histogram,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'mv_boxplot,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'DatasetExperiment_dist,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'DatasetExperiment_boxplot,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'compare_dist,DatasetExperiment'
chart_plot(obj, dobj, eobj)

## S4 method for signature 'DatasetExperiment_heatmap,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'DatasetExperiment_factor_boxplot,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'feature_profile_array,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'feature_profile,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'fold_change_plot,fold_change'
chart_plot(obj, dobj)

## S4 method for signature 'fs_line,forward_selection_by_rank'
chart_plot(obj, dobj)

## S4 method for signature 'glog_opt_plot,glog_transform'
chart_plot(obj, dobj, gobj)

## S4 method for signature 'gs_line,grid_search_1d'
chart_plot(obj, dobj)

## S4 method for signature 'hca_dendrogram,HCA'
chart_plot(obj, dobj)

## S4 method for signature 'kfoldxcv_grid,kfold_xval'
chart_plot(obj, dobj)

## S4 method for signature 'kfoldxcv_metric,kfold_xval'
chart_plot(obj, dobj)

## S4 method for signature 'kw_p_hist,kw_rank_sum'
chart_plot(obj, dobj)

## S4 method for signature 'mv_feature_filter_hist,mv_feature_filter'
chart_plot(obj, dobj)

## S4 method for signature 'mv_sample_filter_hist,mv_sample_filter'
chart_plot(obj, dobj)

## S4 method for signature 'permutation_test_plot,permutation_test'
chart_plot(obj, dobj)

## S4 method for signature 'plsda_feature_importance_plot,PLSDA'
chart_plot(obj, dobj)

## S4 method for signature 'pqn_norm_hist,pqn_norm'
chart_plot(obj, dobj)

## S4 method for signature 'resample_chart,resample'
chart_plot(obj, dobj)

## S4 method for signature 'rsd_filter_hist,rsd_filter'
chart_plot(obj, dobj)

## S4 method for signature 'feature_profile,sb_corr'
chart_plot(obj, dobj, gobj)

## S4 method for signature 'svm_plot_2d,SVM'
chart_plot(obj, dobj, gobj)

## S4 method for signature 'tSNE_scatter,tSNE'
chart_plot(obj, dobj)

## S4 method for signature 'tic_chart,DatasetExperiment'
chart_plot(obj, dobj)

## S4 method for signature 'wilcox_p_hist,wilcox_test'
chart_plot(obj, dobj)

Arguments

Value

a plot object

Examples

C = example_chart()
chart_plot(C,iris_DatasetExperiment())

Description

In univariate classical least squares regression a line is fitted between each feature/variable and a response variable. The fitted line minimises the sum of squared differences between the true response and the predicted response. The coefficients (offset, gradient) of the fit can be tested for significance.

Usage

classical_lsq(alpha = 0.05, mtc = "fdr", factor_names, intercept = TRUE, ...)

Arguments

Value

A classical_lsq object with the following output slots:

Inheritance

A classical_lsq object inherits the following struct classes:

⁠[classical_lsq]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = classical_lsq(
      alpha = 0.05,
      mtc = "fdr",
      factor_names = "V1",
      intercept = FALSE)

D = iris_DatasetExperiment()
M = classical_lsq(factor_names = 'Species')
M = model_apply(M,D)

Description

Histograms and boxplots computed across samples and features are used to visually compare two datasets e.g. before and after filtering and/or normalisation.

Usage

compare_dist(factor_name, ...)

Arguments

Value

A compare_dist object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A compare_dist object inherits the following struct classes:

⁠[compare_dist]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = compare_dist(
      factor_name = "V1")

D1=MTBLS79_DatasetExperiment(filtered=FALSE)
D2=MTBLS79_DatasetExperiment(filtered=TRUE)
C = compare_dist(factor_name='Class')
chart_plot(C,D1,D2)

Description

Univariate least squares regression models are used to compare models with and without potential confounding factors included. The change in coefficients (delta) is then computed for each potential confounding factor. Factors with a large delta are said to be having a large impact on the model and are therefore confounding. p-values are computed for models with confounders included to reduce potential false positives. Only suitable for main factors with 2 levels.

Usage

confounders_clsq(
  alpha = 0.05,
  mtc = "fdr",
  factor_name,
  confounding_factors,
  threshold = 0.15,
  ...
)

Arguments

Value

A confounders_clsq object with the following output slots:

Inheritance

A confounders_clsq object inherits the following struct classes:

⁠[confounders_clsq]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = confounders_clsq(
      alpha = 0.05,
      mtc = "fdr",
      factor_name = character(0),
      confounding_factors = character(0),
      threshold = 0.15)

D = MTBLS79_DatasetExperiment()
M = filter_by_name(mode='include',dimension='variable',
        names=colnames(D$data)[1:10]) + # first 10 features
    filter_smeta(mode='exclude',levels='QC',
        factor_name='Class') + # reduce to two group comparison
    confounders_clsq(factor_name = 'Class',
        confounding_factors=c('run_order','Batch'))
M = model_apply(M,D)

Description

A barchart of the relative change (delta) in regression coefficient when potential confounding factors are included, and excluded, from the model. Factors with a large delta are considered to be confounding factors.

Usage

confounders_lsq_barchart(feature_to_plot, threshold = 10, ...)

Arguments

Value

A confounders_lsq_barchart object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A confounders_lsq_barchart object inherits the following struct classes:

⁠[confounders_lsq_barchart]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = confounders_lsq_barchart(
      feature_to_plot = 1,
      threshold = 10)

D = MTBLS79_DatasetExperiment()
M = filter_by_name(mode='include',dimension='variable',
        names=colnames(D$data)[1:10]) + # first 10 features
    filter_smeta(mode='exclude',levels='QC',
        factor_name='Class') + # reduce to two group comparison
    confounders_clsq(factor_name = 'Class',
        confounding_factors=c('run_order','Batch'))
M = model_apply(M,D)
C = C=confounders_lsq_barchart(feature_to_plot=1,threshold=15)
chart_plot(C,M[3])

Description

A boxplot of the relative change (delta) in regression coefficient when potential confounding factors are included, and excluded, from the model. Factors with a large delta are considered to be confounding factors.

Usage

confounders_lsq_boxplot(threshold = 10, ...)

Arguments

Value

A confounders_lsq_boxplot object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A confounders_lsq_boxplot object inherits the following struct classes:

⁠[confounders_lsq_boxplot]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = confounders_lsq_boxplot(
      threshold = 10)

D = MTBLS79_DatasetExperiment()
M = filter_by_name(mode='include',dimension='variable',
        names=colnames(D$data)[1:10]) + # first 10 features
    filter_smeta(mode='exclude',levels='QC',
        factor_name='Class') + # reduce to two group comparison
    confounders_clsq(factor_name = 'Class',
        confounding_factors=c('run_order','Batch'))
M = model_apply(M,D)
C = C=confounders_lsq_boxplot(threshold=15)
chart_plot(C,M[3])

Description

Each sample is normalised such that the total signal is equal to one (or a scaling factor if specified).

Usage

constant_sum_norm(scaling_factor = 1, ...)

Arguments

Value

A constant_sum_norm object with the following output slots:

Inheritance

A constant_sum_norm object inherits the following struct classes:

⁠[constant_sum_norm]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = constant_sum_norm(
      scaling_factor = 1)

M = constant_sum_norm()

Description

The correlation between features and a set of continuous factor are calculated. Multiple-test corrected p-values are used to indicate whether the computed coefficients may have occurred by chance.

Usage

corr_coef(alpha = 0.05, mtc = "fdr", factor_names, method = "spearman", ...)

Arguments

Details

This object makes use of functionality from the following packages:

• stats

Value

A corr_coef object with the following output slots:

Inheritance

A corr_coef object inherits the following struct classes:

⁠[corr_coef]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

R Core Team (2024). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.

Examples

M = corr_coef(
      alpha = 0.05,
      mtc = "fdr",
      factor_names = "V1",
      method = "spearman")

D = MTBLS79_DatasetExperiment(filtered=TRUE)

# subset for this example
D = D[,1:10]

# convert to numeric for this example
D$sample_meta$sample_order=as.numeric(D$sample_meta$run_order)
D$sample_meta$sample_rep=as.numeric(D$sample_meta$Sample_Rep)

M = corr_coef(factor_names=c('sample_order','sample_rep'))
M = model_apply(M,D)

Description

A boxplot to visualise the distribution of values within a subset of features.

Usage

DatasetExperiment_boxplot(
  factor_name,
  by_sample = TRUE,
  per_class = TRUE,
  number = 50,
  ...
)

Arguments

Value

A DatasetExperiment_boxplot object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

struct object

Inheritance

A DatasetExperiment_boxplot object inherits the following struct classes:

⁠[DatasetExperiment_boxplot]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = DatasetExperiment_boxplot(
      factor_name = "V1",
      by_sample = FALSE,
      per_class = FALSE,
      number = 50)

D = MTBLS79_DatasetExperiment()
C = DatasetExperiment_boxplot(factor_name='Class',number=10,per_class=FALSE)
chart_plot(C,D)

Description

A histogram to visualise the distribution of values within features.

Usage

DatasetExperiment_dist(factor_name, per_class = TRUE, ...)

Arguments

Value

A DatasetExperiment_dist object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A DatasetExperiment_dist object inherits the following struct classes:

⁠[DatasetExperiment_dist]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = DatasetExperiment_dist(
      factor_name = "V1",
      per_class = FALSE)

D = MTBLS79_DatasetExperiment()
C = DatasetExperiment_dist(factor_name='Class')
chart_plot(C,D)

Description

Boxplot for a feature to visualise the distribution of values within each group

Usage

DatasetExperiment_factor_boxplot(feature_to_plot, factor_names, ...)

Arguments

Value

A DatasetExperiment_factor_boxplot object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A DatasetExperiment_factor_boxplot object inherits the following struct classes:

⁠[DatasetExperiment_factor_boxplot]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = DatasetExperiment_factor_boxplot(
      factor_names = "V1",
      feature_to_plot = "V1")

D = iris_DatasetExperiment()
C = DatasetExperiment_factor_boxplot(factor_names='Species',feature_to_plot='Petal.Width')
chart_plot(C,D)

Description

A heatmap to visualise the measured values in a data matrix.

Usage

DatasetExperiment_heatmap(na_colour = "#FF00E4", ...)

Arguments

Details

This object makes use of functionality from the following packages:

• reshape2

Value

A DatasetExperiment_heatmap object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A DatasetExperiment_heatmap object inherits the following struct classes:

⁠[DatasetExperiment_heatmap]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

References

Wickham H (2007). "Reshaping Data with the reshape Package." Journal of Statistical Software, 21(12), 1-20. http://www.jstatsoft.org/v21/i12/.

Examples

M = DatasetExperiment_heatmap(
      na_colour = "#FF00E4")

D = iris_DatasetExperiment()
C = DatasetExperiment_heatmap()
chart_plot(C,D)

Description

Discriminant Factor Analysis (DFA) is a supervised classification method. Using a linear combination of the input variables, DFA finds new orthogonal axes (canonical values) to minimize the variance within each given class and maximize variance between classes.

Usage

DFA(factor_name, number_components = 2, ...)

Arguments

Value

A DFA object with the following output slots:

Inheritance

A DFA object inherits the following struct classes:

⁠[DFA]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Manly B (1986). Multivariate Statistical Methods: A Primer. Chapman and Hall, Boca Raton.

Examples

M = DFA(
      factor_name = "V1",
      number_components = 2)

D = iris_DatasetExperiment()
M = DFA(factor_name='Species')
M = model_apply(M,D)

Description

A scatter plot of the selected DFA components.

Usage

dfa_scores_plot(
  components = c(1, 2),
  points_to_label = "none",
  factor_name,
  ellipse = "all",
  label_filter = character(0),
  label_factor = "rownames",
  label_size = 3.88,
  ...
)

Arguments

Details

This object makes use of functionality from the following packages:

• scales

• ggplot2

Value

A dfa_scores_plot object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A dfa_scores_plot object inherits the following struct classes:

⁠[dfa_scores_plot]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

References

Wickham H, Pedersen T, Seidel D (2023). scales: Scale Functions for Visualization. R package version 1.3.0, https://CRAN.R-project.org/package=scales.

Wickham H (2016). ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York. ISBN 978-3-319-24277-4, https://ggplot2.tidyverse.org.

Examples

M = dfa_scores_plot(
      components = c(1, 2),
      points_to_label = "none",
      factor_name = "V1",
      ellipse = "all",
      label_filter = character(0),
      label_factor = "rownames",
      label_size = 3.88)

D = iris_DatasetExperiment()
M = mean_centre() + DFA(factor_name='Species')
M = model_apply(M,D)
C = dfa_scores_plot(factor_name = 'Species')
chart_plot(C,M[2])

Description

The dispersion ratio (d-ratio) compares the standard deviation (or non-parametric equivalent) of the Quality Control (QC) samples relative to the standard deviation (or non-parametric equivalent) of the samples for each feature. If the d-ratio is greater than a predefined threshold then the observed sample variance could be due to technical variance and the feature is removed.

Usage

dratio_filter(
  threshold = 20,
  qc_label = "QC",
  factor_name,
  method = "ratio",
  dispersion = "sd",
  ...
)

Arguments

Value

A dratio_filter object with the following output slots:

Inheritance

A dratio_filter object inherits the following struct classes:

⁠[dratio_filter]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Broadhurst D, Goodacre R, Reinke SN, Kuligowski J, Wilson ID, Lewis MR, Dunn WB (2018). "Guidelines and considerations for the use of system suitability and quality control samples in mass spectrometry assays applied in untargeted clinical metabolomic studies." Metabolomics, 14(6).

Examples

M = dratio_filter(
      threshold = 20,
      qc_label = "QC",
      factor_name = "V1",
      method = "ratio",
      dispersion = "sd")

D = MTBLS79_DatasetExperiment()
M = dratio_filter(threshold=20,qc_label='QC',factor_name='Class')
M = model_apply(M,D)

Description

Samples are randomly chosen from each level such that the training set has equal numbers of samples for all levels. The number of samples is based on the input proportion and the smallest group size.

Usage

equal_split(p_train = 1, factor_name, ...)

Arguments

Value

A equal_split object with the following output slots:

Inheritance

A equal_split object inherits the following struct classes:

⁠[equal_split]⁠ >> ⁠[split_data]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = equal_split(
      factor_name = "V1",
      p_train = 0.75)

D = iris_DatasetExperiment()
M = equal_split(factor_name='Species')
M = model_apply(M,D)

Description

A boxplot to visualise the distribution of values within a feature.

Usage

feature_boxplot(
  label_outliers = TRUE,
  feature_to_plot,
  factor_name,
  show_counts = TRUE,
  style = "boxplot",
  jitter = FALSE,
  fill = FALSE,
  ...
)

Arguments

Value

A feature_boxplot object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A feature_boxplot object inherits the following struct classes:

⁠[feature_boxplot]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = feature_boxplot(
      label_outliers = FALSE,
      feature_to_plot = "V1",
      factor_name = "V1",
      show_counts = FALSE,
      style = "boxplot",
      jitter = FALSE,
      fill = FALSE)

D = MTBLS79_DatasetExperiment
C = feature_boxplot(factor_name='Species',feature_to_plot='Petal.Width')
chart_plot(C,D)

Description

A plot visualising the change in intensity of a feature with a continuous variable such as time, dose, or run order.

Usage

feature_profile(
  run_order,
  qc_label,
  qc_column,
  colour_by,
  feature_to_plot,
  plot_sd = FALSE,
  ...
)

Arguments

Value

A feature_profile object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A feature_profile object inherits the following struct classes:

⁠[feature_profile]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = feature_profile(
      run_order = character(0),
      qc_label = character(0),
      qc_column = character(0),
      colour_by = character(0),
      feature_to_plot = numeric(0),
      plot_sd = FALSE)

D = MTBLS79_DatasetExperiment()
C = feature_profile(run_order='run_order',
    qc_label='QC',
    qc_column='Class',
    colour_by='Class',
    feature_to_plot=1)
chart_plot(C,D)

Description

A plot visualising the change in intensity of a feature with a continuous variable such as time, dose, or run order.

Usage

feature_profile_array(
  run_order,
  qc_label,
  qc_column,
  colour_by,
  feature_to_plot,
  nrow = 5,
  log = TRUE,
  ...
)

Arguments

Value

A feature_profile_array object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A feature_profile_array object inherits the following struct classes:

⁠[feature_profile_array]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = feature_profile_array(
      run_order = character(0),
      qc_label = character(0),
      qc_column = character(0),
      colour_by = character(0),
      feature_to_plot = numeric(0),
      nrow = 1,
      log = FALSE)

D = MTBLS79_DatasetExperiment()
C = feature_profile_array(
    run_order='run_order',
    qc_label='QC',
    qc_column='Class',
    colour_by='Class',
    feature_to_plot=1:3,
    nrow=1,
    log=TRUE)
chart_plot(C,D)

Description

Filter samples/variables by row/column name, index or logicals.

Usage

filter_by_name(mode = "exclude", dimension = "sample", names, ...)

Arguments

Value

A filter_by_name object with the following output slots:

Inheritance

A filter_by_name object inherits the following struct classes:

⁠[filter_by_name]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = filter_by_name(
      mode = "exclude",
      dimension = "sample",
      names = character(0))

D = MTBLS79_DatasetExperiment()
M = filter_by_name(mode='exclude',dimension='variable',names=c(1,2,3))
M = model_apply(M,D)

Description

The number of measured values is counted for each feature, and any feature with less than a predefined minimum number of values in each group is removed. If there are several factors, then the threshold is applied so that the minimum number of samples is present for all combinations (interactions) of groups.

Usage

filter_na_count(threshold, factor_name, ...)

Arguments

Value

A filter_na_count object with the following output slots:

Inheritance

A filter_na_count object inherits the following struct classes:

⁠[filter_na_count]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = filter_na_count(
      threshold = 2,
      factor_name = "V1")

D = MTBLS79_DatasetExperiment()
M = filter_na_count(threshold=3,factor_name='Class')
M = model_apply(M,D)

Description

The data is filtered by so that the named levels of a factor are included/excluded from the dataset.

Usage

filter_smeta(mode = "include", levels, factor_name, ...)

Arguments

Value

A filter_smeta object with the following output slots:

Inheritance

A filter_smeta object inherits the following struct classes:

⁠[filter_smeta]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = filter_smeta(
      mode = "include",
      levels = character(0),
      factor_name = "V1")

D = MTBLS79_DatasetExperiment()
M = filter_smeta(mode='exclude',levels='QC',factor_name='QC')
M = model_apply(M,D)

Description

A fisher exact test is used to analyse contingency tables by comparing the number of correctly/incorrectly predicted group labels. A multiple test corrected p-value indicates whether the number of measured values is significantly different between groups.

Usage

fisher_exact(alpha = 0.05, mtc = "fdr", factor_name, factor_pred, ...)

Arguments

Value

A fisher_exact object with the following output slots:

Inheritance

A fisher_exact object inherits the following struct classes:

⁠[fisher_exact]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = fisher_exact(
      alpha = 0.05,
      mtc = "fdr",
      factor_name = "V1",
      factor_pred = data.frame(id=NA))

# load some data
D=MTBLS79_DatasetExperiment()

# prepare predictions based on NA
pred=as.data.frame(is.na(D$data))
pred=lapply(pred,factor,levels=c(TRUE,FALSE))
pred=as.data.frame(pred)

# apply method
M = fisher_exact(alpha=0.05,mtc='fdr',factor_name='Class',factor_pred=pred)
M=model_apply(M,D)

Description

Fold change is the relative change in mean (or non-parametric equivalent) intensities of a feature between all pairs of levels in a factor.

Usage

fold_change(
  factor_name,
  paired = FALSE,
  sample_name = character(0),
  threshold = 2,
  control_group = character(0),
  method = "geometric",
  conf_level = 0.95,
  ...
)

Arguments

Value

A fold_change object with the following output slots:

Inheritance

A fold_change object inherits the following struct classes:

⁠[fold_change]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Price Jr RM, Bonett DG (2020). "Confidence Intervals for Ratios of Means and Medians." Journal of Educational and Behavioral Statistics, 45(6), 750-770.

Examples

M = fold_change(
      factor_name = "V1",
      sample_name = character(0),
      paired = FALSE,
      threshold = 2,
      control_group = character(0),
      method = "geometric",
      conf_level = 0.95)

D = MTBLS79_DatasetExperiment()
M = fold_change(factor_name='Class')
M = model_apply(M,D)

Description

For more than one factor the fold change calculation is extended to include all combinations of levels (interactions) of all factors. Paired fold changes are not possible for this computation.

Usage

fold_change_int(
  factor_name,
  threshold = 2,
  control_group = character(0),
  method = "geometric",
  conf_level = 0.95,
  ...
)

Arguments

Value

A fold_change_int object with the following output slots:

Inheritance

A fold_change_int object inherits the following struct classes:

⁠[fold_change_int]⁠ >> ⁠[fold_change]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Lloyd GR, Jankevics A, Weber RJM (2020). "struct: an R/Bioconductor-based framework for standardized metabolomics data analysis and beyond." Bioinformatics, 36(22-23), 5551-5552. https://doi.org/10.1093/bioinformatics/btaa1031.

Examples

M = fold_change_int(
      factor_name = "V1",
      sample_name = character(0),
      threshold = 2,
      control_group = character(0),
      method = "geometric",
      paired = FALSE,
      conf_level = 0.95)

D = MTBLS79_DatasetExperiment()
D=D[,1:10,drop=FALSE]
M = filter_smeta(mode='exclude',levels='QC',factor_name='Class') +
    fold_change_int(factor_name=c('Class','Batch'))
M = model_apply(M,D)

Description

A plot of fold changes calculated for a chosen subset of features. A predefined fold change threshold is indicated by shaded regions.

Usage

fold_change_plot(number_features = 20, orientation = "portrait", ...)

Arguments

Value

A fold_change_plot object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A fold_change_plot object inherits the following struct classes:

⁠[fold_change_plot]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = fold_change_plot(
      number_features = 10,
      orientation = "portrait")

C = fold_change_plot()

Description

A model is trained and performance metric computed by including increasing numbers of features in the model. The features to be included in each step are defined by their rank, which is computed from another variable e.g. VIP score. An "optimal"subset of features is suggested by minimising the input performance metric.

Usage

forward_selection_by_rank(
  min_no_vars = 1,
  max_no_vars = 100,
  step_size = 1,
  factor_name,
  variable_rank,
  ...
)

Arguments

Value

A forward_selection_by_rank object with the following output slots:

Inheritance

A forward_selection_by_rank object inherits the following struct classes:

⁠[forward_selection_by_rank]⁠ >> ⁠[resampler]⁠ >> ⁠[iterator]⁠ >> ⁠[struct_class]⁠

Examples

M = forward_selection_by_rank(
      min_no_vars = 1,
      max_no_vars = 100,
      step_size = 1,
      factor_name = "V1",
      variable_rank = 1)

# some data
D = MTBLS79_DatasetExperiment(filtered=TRUE)

# normalise, impute and scale then remove QCs
P = pqn_norm(qc_label='QC',factor_name='Class') +
    knn_impute(neighbours=5) +
    glog_transform(qc_label='QC',factor_name='Class') +
    filter_smeta(mode='exclude',levels='QC',factor_name='Class')
P = model_apply(P,D)
D = predicted(P)

# forward selection using a PLSDA model
M = forward_selection_by_rank(factor_name='Class',
                             min_no_vars=2,
                             max_no_vars=11,
                             variable_rank=1:2063) *
    (mean_centre() + PLSDA(number_components=1,
                           factor_name='Class'))
M = run(M,D,balanced_accuracy())

Description

A line plot for forward selection. The computed model performance metric is plotted against the number of features included in the model.

Usage

fs_line(...)

Arguments

Value

A fs_line object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A fs_line object inherits the following struct classes:

⁠[fs_line]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = fs_line()

# some data
D = MTBLS79_DatasetExperiment(filtered=TRUE)

# normalise, impute and scale then remove QCs
P = pqn_norm(qc_label='QC',factor_name='Class') +
    knn_impute(neighbours=5) +
    glog_transform(qc_label='QC',factor_name='Class') +
    filter_smeta(mode='exclude',levels='QC',factor_name='Class')
P = model_apply(P,D)
D = predicted(P)

# forward selection using a PLSDA model
M = forward_selection_by_rank(factor_name='Class',
                             min_no_vars=2,
                             max_no_vars=11,
                             variable_rank=1:2063) *
    (mean_centre() + PLSDA(number_components=1,
                           factor_name='Class'))
M = run(M,D,balanced_accuracy())

# chart
C = fs_line()
chart_plot(C,M)

Description

A plot of the sum of squares error (SSE) vs different values of lambda for the glog transform. The indicated optimum value for lambda minimises the SSE.

Usage

glog_opt_plot(plot_grid = 100, ...)

Arguments

Details

This object makes use of functionality from the following packages:

• pmp

Value

A glog_opt_plot object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A glog_opt_plot object inherits the following struct classes:

⁠[glog_opt_plot]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

References

Jankevics A, Lloyd GR, Weber RJM (????). pmp: Peak Matrix Processing and signal batch correction for metabolomics datasets. R package version 1.15.1.

Examples

M = glog_opt_plot(
      plot_grid = numeric(0))

D = iris_DatasetExperiment()
M = glog_transform(qc_label='versicolor',factor_name='Species')
M = model_apply(M,D)
C = glog_opt_plot()
chart_plot(C,M,D)

Description

The generalised logarithm (glog) transformation applies a log transformation while applying an offset to account for technical variation.

Usage

glog_transform(qc_label = "QC", factor_name, lambda = NULL, ...)

Arguments

Details

This object makes use of functionality from the following packages:

• pmp

Value

A glog_transform object with the following output slots:

Inheritance

A glog_transform object inherits the following struct classes:

⁠[glog_transform]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Jankevics A, Lloyd GR, Weber RJM (????). pmp: Peak Matrix Processing and signal batch correction for metabolomics datasets. R package version 1.15.1.

Durbin B, Hardin J, Hawkins D, Rocke D (2002). "A variance-stabilizing transformation for gene-expression microarray data." Bioinformatics, 18(Suppl 1), S105-S110.

Parsons HM, Ludwig C, Gunther UL, Viant MR (2007). "Improved classification accuracy in 1- and ', '2-dimensional NMR metabolomics data using the variance ', 'stabilising generalised logarithm transformation." Bioinformatics, 8(1), 234.

Examples

M = glog_transform(
      qc_label = "QC",
      factor_name = "V1",
      lambda = NULL)

D = iris_DatasetExperiment()
M = glog_transform(qc_label='versicolor',factor_name='Species')
M = model_apply(M,D)

Description

A one dimensional grid search calculates a performance metric for a model at evenly spaced values for a model input parameter. The "optimum" value for the parameter is suggested as the one which maximises performance, or minimises error (whichever is appropriate for the chosen metric)

Usage

grid_search_1d(
  param_to_optimise,
  search_values,
  model_index,
  factor_name,
  max_min = "min",
  ...
)

Arguments

Value

A grid_search_1d object with the following output slots:

Inheritance

A grid_search_1d object inherits the following struct classes:

⁠[grid_search_1d]⁠ >> ⁠[resampler]⁠ >> ⁠[iterator]⁠ >> ⁠[struct_class]⁠

Examples

M = grid_search_1d(
      param_to_optimise = character(0),
      search_values = numeric(0),
      model_index = numeric(0),
      factor_name = "V1",
      max_min = "min")

D = MTBLS79_DatasetExperiment()
# some preprocessing
M = pqn_norm(qc_label='QC',factor_name='Class') +
    knn_impute() +
    glog_transform(qc_label='QC',factor_name='Class') +
    filter_smeta(factor_name='Class',levels='QC',mode='exclude')
M=model_apply(M,D)
D=predicted(M)

# reduce number of features for this example
D=D[,1:10]

# optmise number of components for PLS model
I = grid_search_1d(param_to_optimise='number_components',search_values=1:5,
        model_index=2,factor_name='Class') *
        (mean_centre()+PLSDA(factor_name='Class'))
I = run(I,D,balanced_accuracy())

Description

A plot of the calculated performance metric against the model input parameter values used to train the model. The optimum parameter value is indicated based on minimising (or maximising) the chosen metric.

Usage

gs_line(...)

Arguments

Value

A gs_line object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A gs_line object inherits the following struct classes:

⁠[gs_line]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = gs_line()

C = gs_line()

Description

Hierarchical Cluster Analysis is a numerical technique that uses agglomerative clustering to identify clusters or groupings of samples.

Usage

HCA(
  dist_method = "euclidean",
  cluster_method = "complete",
  minkowski_power = 2,
  factor_name,
  ...
)

Arguments

Details

This object makes use of functionality from the following packages:

• stats

Value

A HCA object with the following output slots:

Inheritance

A HCA object inherits the following struct classes:

⁠[HCA]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

R Core Team (2024). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.

Examples

M = HCA(
      dist_method = "euclidean",
      cluster_method = "complete",
      minkowski_power = numeric(0),
      factor_name = "V1")

D = iris_DatasetExperiment()
M = HCA(factor_name='Species')
M = model_apply(M,D)

Description

A dendrogram visualising the clustering by HCA.

Usage

hca_dendrogram(...)

Arguments

Details

This object makes use of functionality from the following packages:

• ggdendro

Value

A hca_dendrogram object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A hca_dendrogram object inherits the following struct classes:

⁠[hca_dendrogram]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

References

de Vries A, Ripley BD (2024). ggdendro: Create Dendrograms and Tree Diagrams Using 'ggplot2'. R package version 0.2.0, https://CRAN.R-project.org/package=ggdendro.

Examples

M = hca_dendrogram()

C = hca_dendrogram()

Description

Tukey's HSD post hoc test is a modified t-test applied for all features to all pairs of levels in a factor. It is used to determine which groups are different (if any). A multiple test corrected p-value is computed to indicate which groups are significantly different to the others for each feature.

Usage

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

Arguments

Details

This object makes use of functionality from the following packages:

• agricolae

Value

A HSD object with the following output slots:

Inheritance

A HSD object inherits the following struct classes:

⁠[HSD]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

de Mendiburu F (2023). agricolae: Statistical Procedures for Agricultural Research. R package version 1.3-7, https://CRAN.R-project.org/package=agricolae.

Examples

M = HSD(
      alpha = 0.05,
      mtc = "fdr",
      formula = y ~ x,
      unbalanced = FALSE)

D = iris_DatasetExperiment()
M = HSD(formula=y~Species)
M = model_apply(M,D)

Description

Tukey's HSD post hoc test is a modified t-test applied for all features to all pairs of levels in a factor. It is used to determine which groups are different (if any). A multiple test corrected p-value is computed to indicate which groups are significantly different to the others for each feature. For mixed effects models estimated marginal means are used.

Usage

HSDEM(alpha = 0.05, mtc = "fdr", formula, ...)

Arguments

Details

This object makes use of functionality from the following packages:

• emmeans

• nlme

Value

A HSDEM object with the following output slots:

Inheritance

A HSDEM object inherits the following struct classes:

⁠[HSDEM]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Lenth R (2024). emmeans: Estimated Marginal Means, aka Least-Squares Means. R package version 1.10.1, https://CRAN.R-project.org/package=emmeans.

Pinheiro J, Bates D, R Core Team (2023). nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1-164, https://CRAN.R-project.org/package=nlme.

Pinheiro JC, Bates DM (2000). Mixed-Effects Models in S and S-PLUS. Springer, New York. doi:10.1007/b98882 https://doi.org/10.1007/b98882.

Examples

M = HSDEM(
      alpha = 0.05,
      mtc = "fdr",
      formula = y ~ x)

D = iris_DatasetExperiment()
D$sample_meta$id=rownames(D) # dummy id column
M = HSDEM(formula = y~Species+ Error(id/Species))
M = model_apply(M,D)

Description

k-fold cross-validation is an iterative approach applied to validate models. The samples are divided into k "folds", or subsets. Each subset is excluded from model training and used for model validation once, resulting in a single left-out prediction for each sample. Model performance metrics are then computed for the training and test sets across all folds.

Usage

kfold_xval(folds = 10, method = "venetian", factor_name, collect = NULL, ...)

Arguments

Value

A kfold_xval object with the following output slots:

Inheritance

A kfold_xval object inherits the following struct classes:

⁠[kfold_xval]⁠ >> ⁠[resampler]⁠ >> ⁠[iterator]⁠ >> ⁠[struct_class]⁠

Examples

M = kfold_xval(
      folds = 5,
      method = "random",
      factor_name = "V1",
      collect = NULL)

D = iris_DatasetExperiment()
I = kfold_xval(factor_name='Species') *
    (mean_centre() + PLSDA(factor_name='Species'))
I = run(I,D,balanced_accuracy())

Description

A graphic for visualising the true class and the predicted class of samples in all groups for all cross-validation folds.

Usage

kfoldxcv_grid(factor_name, level, ...)

Arguments

Value

A kfoldxcv_grid object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A kfoldxcv_grid object inherits the following struct classes:

⁠[kfoldxcv_grid]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = kfoldxcv_grid(
      factor_name = "V1",
      level = "level_1")

D = iris_DatasetExperiment()
I = kfold_xval(factor_name='Species') *
    (mean_centre() + PLSDA(factor_name='Species'))
I = run(I,D,balanced_accuracy())

C = kfoldxcv_grid(factor_name='Species',level='setosa')
chart_plot(C,I)

Description

A boxplot of the performance metric computed for each fold of a k-fold cross-validation.

Usage

kfoldxcv_metric(...)

Arguments

Value

A kfoldxcv_metric object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A kfoldxcv_metric object inherits the following struct classes:

⁠[kfoldxcv_metric]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = kfoldxcv_metric()

C = kfoldxcv_metric()

Description

k-nearest neighbour missing value imputation replaces missing values in the data with the average of a predefined number of the most similar neighbours for which the value is present

Usage

knn_impute(
  neighbours = 5,
  sample_max = 50,
  feature_max = 50,
  by = "features",
  ...
)

Arguments

Details

This object makes use of functionality from the following packages:

• pmp

Value

A knn_impute object with the following output slots:

Inheritance

A knn_impute object inherits the following struct classes:

⁠[knn_impute]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Jankevics A, Lloyd GR, Weber RJM (????). pmp: Peak Matrix Processing and signal batch correction for metabolomics datasets. R package version 1.15.1.

Examples

M = knn_impute(
      neighbours = 5,
      feature_max = 50,
      sample_max = 50,
      by = "features")

M = knn_impute()

Description

A histogram of the p-values computed by the kruskal-wallis method

Usage

kw_p_hist(...)

Arguments

Value

A kw_p_hist object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A kw_p_hist object inherits the following struct classes:

⁠[kw_p_hist]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = kw_p_hist()

C = kw_p_hist()

Description

The Kruskal-Wallis test is a univariate hypothesis testing method that allows multiple (n>=2) groups to be compared without making the assumption that values are normally distributed. It is the non-parametric equivalent of a 1-way ANOVA. The test is applied to all variables/features individually, and multiple test corrected p-values are computed to indicate the significance of variables/features.

Usage

kw_rank_sum(alpha = 0.05, mtc = "fdr", factor_names, ...)

Arguments

Value

A kw_rank_sum object with the following output slots:

Inheritance

A kw_rank_sum object inherits the following struct classes:

⁠[kw_rank_sum]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = kw_rank_sum(
      alpha = 0.05,
      mtc = "fdr",
      factor_names = "V1")

D = iris_DatasetExperiment()
M = kw_rank_sum(factor_names='Species')
M = model_apply(M,D)

Description

Linear models can be used to carry out regression, single stratum analysis of variance and analysis of covariance.

Usage

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

Arguments

Details

This object makes use of functionality from the following packages:

• stats

Value

A linear_model object with the following output slots:

Inheritance

A linear_model object inherits the following struct classes:

⁠[linear_model]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

R Core Team (2024). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.

Examples

M = linear_model(
      formula = y ~ x,
      na_action = "na.omit",
      contrasts = list())

D = iris_DatasetExperiment()
M = linear_model(formula = y~Species)

Description

A logarithmic transform is applied to all values in the data matrix.

Usage

log_transform(base = 10, ...)

Arguments

Value

A log_transform object with the following output slots:

struct object

Inheritance

A log_transform object inherits the following struct classes:

⁠[log_transform]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = log_transform(
      base = 10)

M = log_transform()

Description

The mean sample is subtracted from all samples in the data matrix. The features in the centred matrix all have zero mean.

Usage

mean_centre(mode = "data", ...)

Arguments

Value

A mean_centre object with the following output slots:

Inheritance

A mean_centre object inherits the following struct classes:

⁠[mean_centre]⁠ >> ⁠[preprocess]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = mean_centre(
      mode = "data")

M = mean_centre()

Description

The data matrix is normalised by the mean of the median of each factor level.

Usage

mean_of_medians(factor_name, ...)

Arguments

Value

A mean_of_medians object with the following output slots:

Inheritance

A mean_of_medians object inherits the following struct classes:

⁠[mean_of_medians]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

Examples

M = mean_of_medians(
      factor_name = "V1")

D = iris_DatasetExperiment()
M = mean_of_medians(factor_name='Species')
M = model_apply(M,D)

Description

A mixed effects model is an extension of ANOVA where there are both fixed and random effects.

Usage

mixed_effect(alpha = 0.05, mtc = "fdr", formula, ss_type = "marginal", ...)

Arguments

Details

This object makes use of functionality from the following packages:

• nlme

• emmeans

Value

A mixed_effect object with the following output slots:

Inheritance

A mixed_effect object inherits the following struct classes:

⁠[mixed_effect]⁠ >> ⁠[ANOVA]⁠ >> ⁠[model]⁠ >> ⁠[stato]⁠ >> ⁠[struct_class]⁠

References

Pinheiro J, Bates D, R Core Team (2023). nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1-164, https://CRAN.R-project.org/package=nlme.

Pinheiro JC, Bates DM (2000). Mixed-Effects Models in S and S-PLUS. Springer, New York. doi:10.1007/b98882 https://doi.org/10.1007/b98882.

Lenth R (2024). emmeans: Estimated Marginal Means, aka Least-Squares Means. R package version 1.10.1, https://CRAN.R-project.org/package=emmeans.

Fox J, Weisberg S (2019). An R Companion to Applied Regression, Third edition. Sage, Thousand Oaks CA. https://socialsciences.mcmaster.ca/jfox/Books/Companion/.

Examples

M = mixed_effect(
      alpha = 0.05,
      mtc = "fdr",
      formula = y ~ x,
      ss_type = "marginal")

D = iris_DatasetExperiment()
D$sample_meta$id=rownames(D) # dummy id column
M = mixed_effect(formula = y~Species+ Error(id/Species))
M = model_apply(M,D)

Description

Applies method to the input DatasetExperiment

Usage

## S4 method for signature 'ANOVA,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'HSD,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'mixed_effect,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'HSDEM,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'classical_lsq,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'confounders_clsq,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'constant_sum_norm,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'corr_coef,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'split_data,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'equal_split,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'filter_smeta,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'fisher_exact,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'fold_change,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'fold_change_int,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'HCA,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'knn_impute,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'kw_rank_sum,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'log_transform,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'mean_of_medians,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'nroot_transform,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'pairs_filter,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'prop_na,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'rsd_filter,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'sb_corr,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'stratified_split,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'tSNE,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'ttest,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'vec_norm,DatasetExperiment'
model_apply(M, D)

## S4 method for signature 'wilcox_test,DatasetExperiment'
model_apply(M, D)

Arguments

Value

Returns a modified method object

Examples

M=model()
model_apply(M,DatasetExperiment())

Description

Apply a model using the input DatasetExperiment. Assumes the model is trained first.

Usage

## S4 method for signature 'DFA,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'PCA,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'PLSR,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'PLSDA,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'autoscale,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'blank_filter,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'constant_sum_norm,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'dratio_filter,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'filter_by_name,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'filter_na_count,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'filter_smeta,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'glog_transform,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'linear_model,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'mean_centre,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'mv_feature_filter,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'mv_sample_filter,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'OPLSR,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'OPLSDA,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'pareto_scale,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'pqn_norm,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'SVM,DatasetExperiment'
model_predict(M, D)

## S4 method for signature 'vec_norm,DatasetExperiment'
model_predict(M, D)

Arguments

Value

Returns a modified model object

Examples

M = example_model()
M = model_predict(M,iris_DatasetExperiment())

Description

Reverse the effect of a preprocessing step on a DatasetExperiment.

Usage

## S4 method for signature 'autoscale,DatasetExperiment'
model_reverse(M, D)

## S4 method for signature 'mean_centre,DatasetExperiment'
model_reverse(M, D)

Arguments

Value

Returns a modified DatasetExperiment object

Examples

M = example_model()
D = model_reverse(M,iris_DatasetExperiment())

Description

Trains a model using the input DatasetExperiment

Usage

## S4 method for signature 'DFA,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'PCA,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'PLSR,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'PLSDA,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'autoscale,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'blank_filter,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'constant_sum_norm,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'dratio_filter,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'filter_by_name,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'filter_na_count,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'filter_smeta,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'glog_transform,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'linear_model,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'mean_centre,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'mv_feature_filter,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'mv_sample_filter,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'OPLSR,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'OPLSDA,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'pareto_scale,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'pqn_norm,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'SVM,DatasetExperiment'
model_train(M, D)

## S4 method for signature 'vec_norm,DatasetExperiment'
model_train(M, D)

Arguments

Value

Returns a modified model object

Examples

M = example_model()
M = model_train(M,iris_DatasetExperiment())

Description

Direct-infusion mass spectrometry (DIMS) metabolomics is an important approach for characterising molecular responses of organisms to disease, drugs and the environment. Increasingly large-scale metabolomics studies are being conducted, necessitating improvements in both bioanalytical and computational workflows to maintain data quality. This dataset represents a systematic evaluation of the reproducibility of a multi-batch DIMS metabolomics study of cardiac tissue extracts. It comprises of twenty biological samples (cow vs. sheep) that were analysed repeatedly, in 8 batches across 7 days, together with a concurrent set of quality control (QC) samples. Data are presented from each step of the workflow and are available in MetaboLights (https://www.ebi.ac.uk/metabolights/MTBLS79)

Usage

MTBLS79_DatasetExperiment(filtered = FALSE)

Arguments

Value

DatasetExperiment object

Examples

D = MTBLS79_DatasetExperiment()
summary(D)

Description

Boxplots of the number of missing values per sample/feature.

Usage

mv_boxplot(
  label_outliers = TRUE,
  by_sample = TRUE,
  factor_name,
  show_counts = TRUE,
  ...
)

Arguments

Value

A mv_boxplot object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A mv_boxplot object inherits the following struct classes:

⁠[mv_boxplot]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = mv_boxplot(
      label_outliers = FALSE,
      by_sample = FALSE,
      factor_name = "V1",
      show_counts = FALSE)

D = MTBLS79_DatasetExperiment()
C = mv_boxplot(factor_name='Class')
chart_plot(C,D)

Description

Removes features where the percentage of non-missing values falls below a threshold.

Usage

mv_feature_filter(
  threshold = 20,
  qc_label = "QC",
  method = "QC",
  factor_name,
  ...
)

Arguments

Details

This object makes use of functionality from the following packages:

• pmp

Value

A mv_feature_filter object with the following output slots:

Inheritance

A mv_feature_filter object inherits the following struct classes:

⁠[mv_feature_filter]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Jankevics A, Lloyd GR, Weber RJM (????). pmp: Peak Matrix Processing and signal batch correction for metabolomics datasets. R package version 1.15.1.

Examples

M = mv_feature_filter(
      threshold = 20,
      qc_label = "QC",
      method = "QC",
      factor_name = "V1")

D = iris_DatasetExperiment()
M = mv_feature_filter(factor_name='Species',qc_label='versicolor')
M = model_apply(M,D)

Description

A histogram of the proportion of missing values per feature.

Usage

mv_feature_filter_hist(...)

Arguments

Value

A mv_feature_filter_hist object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A mv_feature_filter_hist object inherits the following struct classes:

⁠[mv_feature_filter_hist]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = mv_feature_filter_hist()

C = mv_feature_filter_hist()

Description

A histogram of the numbers of missing values per sample/feature

Usage

mv_histogram(label_outliers = TRUE, by_sample = TRUE, ...)

Arguments

Value

A mv_histogram object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

struct object

Inheritance

A mv_histogram object inherits the following struct classes:

⁠[mv_histogram]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = mv_histogram(
      label_outliers = FALSE,
      by_sample = FALSE)

D = MTBLS79_DatasetExperiment()
C = mv_histogram(label_outliers=FALSE,by_sample=FALSE)
chart_plot(C,D)

Description

Removes samples where the percent number of missing values exceeds a threshold.

Usage

mv_sample_filter(mv_threshold = 20, ...)

Arguments

Details

This object makes use of functionality from the following packages:

• pmp

Value

A mv_sample_filter object with the following output slots:

Inheritance

A mv_sample_filter object inherits the following struct classes:

⁠[mv_sample_filter]⁠ >> ⁠[model]⁠ >> ⁠[struct_class]⁠

References

Jankevics A, Lloyd GR, Weber RJM (????). pmp: Peak Matrix Processing and signal batch correction for metabolomics datasets. R package version 1.15.1.

Examples

M = mv_sample_filter(
      mv_threshold = 20)

C = mv_sample_filter()

Description

A histogram of the the proportion of missing values per sample

Usage

mv_sample_filter_hist(...)

Arguments

Value

A mv_sample_filter_hist object. This object has no output slots. See chart_plot in the struct package to plot this chart object.

Inheritance

A mv_sample_filter_hist object inherits the following struct classes:

⁠[mv_sample_filter_hist]⁠ >> ⁠[chart]⁠ >> ⁠[struct_class]⁠

Examples

M = mv_sample_filter_hist()

C = mv_sample_filter_hist()