r-matrix-manipulation

Use when when working with three-part metabolomics data structures (featuredata matrix, metabolitedata and sampledata dataframes) and you need to: (1) identify subsets of metabolites by their annotati

SKILL.md

GitHub

R matrix manipulation for metabolomics feature data

Summary

Programmatically load, index, and transform metabolomics feature matrices in R using vectorized operations and conditional subsetting. This skill enables extraction of control metabolite indices, application of normalization workflows, and assembly of multi-component result objects from featuredata, sampledata, and metabolitedata inputs.

When to use

When working with three-part metabolomics data structures (featuredata matrix, metabolitedata and sampledata dataframes) and you need to: (1) identify subsets of metabolites by their annotation (e.g., negative controls via a Boolean column), (2) apply a normalization function that returns a list object and extract normalized featuredata plus computed components (e.g., unwanted variation matrices), or (3) store the results as a named list object for downstream analysis.

When NOT to use

  • Input data is already in a pre-normalized or processed format and control metabolite identification is not needed
  • Metabolite annotations (neg_control or equivalent column) are missing from metabolitedata; the conditional indexing step will fail or return empty indices
  • The normalization function has already been applied and the result object is being reprocessed; re-running risks double-normalization

Inputs

  • featuredata (numeric matrix: samples × metabolites)
  • metabolitedata (dataframe with metabolite annotations, including control status columns)
  • sampledata (dataframe with sample metadata)
  • control indices (integer vector from conditional indexing)

Outputs

  • uv_ruvrandclust or similar named list object containing: normalized featuredata, unwanted variation matrix (uvdata), and associated metadata
  • Extracted normalized featuredata matrix
  • Computed variation components

How to apply

Load the metabolomics dataset (three components: featuredata as a numeric matrix with samples as rows and metabolites as columns; metabolitedata as a dataframe with metabolite names as row names; sampledata as a dataframe with sample names as row names) into R using the NormalizeMets package or native data I/O. Use vectorized conditional indexing (e.g., which(metabolitedata$neg_control==1)) to identify control metabolite column indices based on metadata annotations. Pass the featuredata and control indices to the normalization function (e.g., NormQcmets() with parameters: method='ruvrandclust', k=1, qcmets=control_indices) and capture the returned list object. Extract the normalized featuredata, computed variation components (e.g., uvdata), and metadata from the result list and assign to a new named object (e.g., uv_ruvrandclust <- list(featuredata=..., uvdata=..., metadata=...)). This approach ensures reproducibility and enables chaining of downstream analyses.

Related tools

  • NormalizeMets (Provides NormQcmets() function to apply ruvrandclust normalization on featuredata using control metabolite indices and parameters (method, k, qcmets); returns list object with normalized data and computed components) — github.com/metabolomicstats/NormalizeMets
  • R (Environment for matrix operations, conditional indexing (which()), list construction, and data structure manipulation)
  • RStudio (Recommended IDE for interactive R code execution and data exploration during matrix manipulation workflows)

Examples

neg_control_idx <- which(UVdata$metabolitedata$neg_control==1); uv_ruvrandclust <- NormQcmets(UVdata$featuredata, factors=NULL, method='ruvrandclust', k=1, qcmets=neg_control_idx)

Evaluation signals

  • Verify the control_indices vector has length > 0 and all values are valid column indices in featuredata (1 ≤ index ≤ ncol(featuredata))
  • Check that the returned normalized featuredata has identical dimensions to input featuredata (same number of rows and columns)
  • Confirm that uvdata component (unwanted variation matrix) has the same number of rows as featuredata and a number of columns equal to k or the number of principal components extracted
  • Validate that no NaN, Inf, or unexpected missing values were introduced by the normalization; use summary() and is.na() checks
  • Verify row names (sample identifiers) and column names (metabolite identifiers) are preserved in the normalized featuredata output

Limitations

  • The skill requires that metabolitedata includes a control status column (e.g., neg_control); if annotations are incomplete or inconsistent, indexing will fail or return incomplete subsets.
  • The featuredata matrix must be numeric; non-numeric or mixed-type columns will cause errors during normalization and component extraction.
  • Parameters such as k (number of components) and the choice of qcmets (control metabolites vs. all metabolites) are user-specified and require domain knowledge; incorrect specification can yield biologically misleading normalization.
  • The skill assumes three-part input structure; datasets with alternative formats (e.g., transposed matrices, single merged tables) require reshaping before the workflow can be applied.

Evidence

  • [other] Load the UVdata dataset (featuredata, sampledata, metabolitedata) into R using the NormalizeMets package: "Load the UVdata dataset (featuredata, sampledata, metabolitedata) into R using the NormalizeMets package."
  • [other] Identify negative control metabolites from metabolitedata using the neg_control column (which(UVdata$metabolitedata$neg_control==1)): "Identify negative control metabolites from metabolitedata using the neg_control column (which(UVdata$metabolitedata$neg_control==1))."
  • [other] Apply NormQcmets with method='ruvrandclust', k=1, and qcmets set to the negative control indices to perform remove-unwanted-variation normalization: "Apply NormQcmets with method='ruvrandclust', k=1, and qcmets set to the negative control indices to perform remove-unwanted-variation normalization with clustering on the UVdata featuredata."
  • [other] Extract the normalized featuredata, uvdata (removed unwanted-variation component), and metadata from the returned object and store as uv_ruvrandclust: "Extract the normalized featuredata, uvdata (removed unwanted-variation component), and metadata from the returned object and store as uv_ruvrandclust."
  • [readme] featuredata is the metabolomics data matrix containing all metabolite peak intensities (or concentrations). Unique sample names must be provided as row names and unique metabolite names as column names: "featuredata which is the metabolomics data matrix containing all metabolite peak intensities (or concentrations). Unique sample names must be provided as row names and unique metabolite names as"