Skip to contents

Calculate annotation statistics for all m/z value suggesting the most likely metabolite based on correlated pathway expression.

Source: R/PathwaybasedAnnotationStatistics.R
CalculateAnnotationStatistics.Rd

This function evaluates the potential biological relevance of an annotation for a given m/z value by:

  • Identifying pathways associated with each possible annotated metabolite

  • Calculating colocalisation score between the m/z intensity and the expression of each corresponding pathway

  • Ranking annotations by a combined z-score based on correlation strength and number of supporting significant pathways

  • NOTE: this function requires CreatePathwayAssay to be run first

Usage

CalculateAnnotationStatistics(
  data,
  mz.assay,
  pathway.assay,
  mz.slot = "scale.data",
  pathway.slot = "scale.data",
  return.top = TRUE,
  corr_theshold = 0,
  corr_weight = 1,
  n_weight = 1
)

Arguments

data

A SpaMTP Seurat object containing both metabolite and RAMP_ID assays generated from CreatePathwayAssay.

mz.assay

Character string defining the name of the assay containing m/z features.

pathway.assay

Character string matching the name of the assay containing RAMP_ID features (default = "pathway").

mz.slot

Character string stating the slot to extract m/z values from (default = "scale.data").

pathway.slot

Character string defining the slot to extract pathway features from (default = "scale.data").

return.top

Boolean indicating whether to return only the most likely metabolite with it's corresponding pval and score. If set to FALSE, a list will be returned with statistics for all possible metabolites per m/z (default = TRUE).

corr_theshold

Numeric value stating the correlation threshold to consider a pathway as significantly colocalized. If set to 0, all pathways will be counted (default = 0).

corr_weight

Numeric weight applied to correlation score in z-score calculation. If significance should be based more on the correlation, increase this value (default = 1).

n_weight

Numeric weight applied to number of correlated pathways in z-score calculation (default = 1).

mz

Character or numeric. The target m/z feature. If numeric, the closest matching m/z in the dataset will be selected.

Value

Either a data.frame containing the original annotations for all m/z values and their corresponding most likely metabolite, or a list contating statistics for each m/z value.

Examples

#CalculateAnnotationStatistics(data = data,mz.assay = "SPM",pathway.assay = "merged",mz.slot = "scale.data")