Provides functionality for untargeted LC-MS metabolomics research as specified in the associated publication in the 'Metabolomics Data Processing and Data Analysis—Current Best Practices' special issue of the Metabolites journal (2020). This includes tabular data preprocessing, feature selection and supporting visualizations. Raw data preprocessing and functionality related to biological context, such as pathway analysis, is not included.
Details
In roughly chronological order, the functionality of notame is as follows. Please see the vignettes and paper for more information.
Tabular data preprocessing (reducing unwanted variation and completing the dataset, returning modifed objects):
mark_nasmark specified values as missingflag_detectionflag features with low detection rateflag_contaminantsflag contaminants based on blankscorrect_driftcorrect drift using cubic splineruvs_qcRemove Unwanted Variation (RUV) between batchespca_bhattacharyya_distBhattacharyya distance between batches in PCA spaceperform_repeatabilitycompute repeatability measuresassess_qualityassess quality information of featuresqualityextract quality information of featuresflag_qualityflag low-quality featuresflag_reportreport of flagged featuresimpute_rfimpute missing values using random forestimpute_simpleimpute missing values using simple imputationcluster_featurescluster correlated features originating from the same metabolitecompress_clustersCompress clusters of features to a single featureloglogarithmsexponentialexponentialpqn_normalizationprobabilistic quotient normalizationinverse_normalizeinverse-rank normalizationscalescale
Tabular data preprocessing visualizations (saved to file by default):
visualizationswrite all relevant data preprocessing visualizations to pdfplot_injection_lmestimate the magnitude of driftplot_sample_boxplotsplot a boxplot for each sampleplot_dist_densityplot distance densityplot_tsne_arrows,plot_pca_arrowst-SNE and PCA plots with arrowsplot_tsne_hexbin,plot_pca_hexbint-SNE and PCA hexbin plotsplot_dendrogramsample dendrogramplot_sample_heatmapsample heatmapplot_pca_loadingsPCA loadings plotplot_qualityplot quality metrics
Feature selection – Univariate analysis (return data.frames):
perform_lmlinear modelsperform_logisticlogistic regressionperform_lmerlinear mixed modelsperform_oneway_anovaWelch’s ANOVA and classic ANOVAperform_lm_anovalinear models ANOVA tableperform_t_testpairwise and paired t-testsperform_kruskal_wallisKruskal-Wallis rank-sum testperform_non_parametricpairwise and paired non- parametric testsperform_correlation_testscorrelation testperform_aucarea under curveperform_homoscedasticity_teststest homoscedasticitycohens_dCohen's Dfold_changefold changesummary_statisticssummary statisticssummarize_resultsstatistics cleaning
Feature selection – Supervised learning (return various objects):
muvr_analysismultivariate modelling with minimally biased variable selection (MUVR2)mixomics_pls,mixomics_plsdaa simple PLS(-DA) model with set number of components and all featuresmixomics_pls_optimize,mixomics_plsda_optimizetest different numbers of components for PLS(-DA)mixomics_spls_optimize,mixomics_splsda_optimizetest different numbers of components and features for PLS(-DA)fit_rf,importance_rffit random forest and feature importanceperform_permanovaPERMANOVA
Feature-wise visualizations (these are often drawn for a subset of interesting features after analysis, saved by default):
save_beeswarm_plotssave beeswarm plots of each feature by groupsave_group_boxplotssave box plots of each feature by groupsave_scatter_plotssave scatter plots of each feature against a set variablesave_subject_line_plotssave line plots with meansave_group_lineplotssave line plots with errorbars by groupsave_batch_plotssave batch correction plots
Results visualizations (returned by default, save them using
save_plot):
plot_p_histogramhistogram of p-valuesvolcano_plotvolcano plotmanhattan_plotmanhattan plotmz_rt_plotm/z vs retention time plot (cloud plot)plot_effect_heatmapheatmap of effects between variables, such as correlations
Object utilities:
read_from_excelread formatted Excel filesconstruct_metabosetsconstruct MetaboSet objectswrite_to_excelwrite results to Excel filegroup_colget and set name of the special column for group labelstime_colget and set the name of the special column for time pointssubject_colget and set the name of the special column for subject identifiersflagget and set values in the flag columndrop_flaggeddrop flagged featuresdrop_qcsdrop QC samplesjoin_fDatajoin new columns to feature datajoin_pDatajoin new columns to pheno datacombined_dataretrieve both sample information and featuresmerge_metabosetsmerge MetaboSet objects togethermerge_objectsmerge SummarizedExperiment objects togetherfix_objectfix object for functioning of notameMetaboSet-classMetaboSet
Other utilities:
References
Klåvus et al. (2020). "notame": Workflow for Non-Targeted LC-MS Metabolic Profiling. Metabolites, 10: 135.