filtering-funs: Functions to apply cluster-based filtering
In b2slab/mWISE: Untargeted LC-MS data annotation

Description Usage Arguments Value Examples

Function clusterBased.filter selects only those proposed compounds for which a quasimolecular adduct or fragment has been also proposed in another peak of the same cluster.

Function dataPrep prepares the intensity and retention time data for spectral clustering. Function .LaplacianNg computes a normalized Laplacian matrix.

Function eps.optimization optimizes the epsilon parameter of the dbscan algorithm.

Function featuresClustering performs spectral clustering to group those features that come from the same metabolite. It uses dataPrep, .LaplacianNg, k.optimization and eps.optimization functions. The correlation is computed using the function cor(use = "pairwise.complete.obs").

Function k.optimization optimizes the number of clusters. This value will be used to define the number of eigenvectors considered in the spectral clustering.

Function recoveringPeaks recovers the peaks that have been removed from the first annotated object.

clusterBased.filter(
  df,
  Add.Id = NULL,
  Freq = 0.5,
  Info.Add = NULL,
  polarity,
  do.Par = TRUE,
  nClust = 2
)

dataPrep(IData, Rt, Rt.05 = 5, use = "everything", method = "pearson")

eps.optimization(
  pca.to.tune,
  data.prep,
  IData,
  use = "everything",
  k.tuned,
  method = "pearson",
  do.Par,
  nClust
)

featuresClustering(
  Peak.List,
  Intensity.idx,
  use = "everything",
  method = "pearson",
  Rt.05 = 5,
  do.Par = TRUE,
  nClust
)

k.optimization(
  pca.to.tune,
  data.prep,
  IData,
  nrow.List,
  use = "everything",
  method = "pearson",
  do.Par = TRUE,
  nClust
)

recoveringPeaks(Annotated.Tab, MH.Tab)

`df`	Columns may contain: "Compound", "Add.Id", "Isotope" "Compound" for the proposed candidates. "Add.Id" for the adduct or fragment proposed. "Isotope" to identify the proposed isotopologues.
`Add.Id`	It indicates the adduct(s) or fragment(s) that are required to exist. If NULL, those adducts with an observed frequency equal or higher than 0.50 will be used.
`Freq`	Minimum observed frequency to consider an adduct or a fragment to apply the filter (Def: 0.5).
`Info.Add`	Data frame with adducts and in source fragments information. If NULL, the default mWISE table will be loaded. The columns should be: name with the name of the adduct or fragment nmol with the number of molecules (i.e., 2M+H: nmol=2 ) charge with the charge of the adduct or fragment (i.e., M+3H: charge=3) massdiff with the mass difference (i.e., M+H: massdiff=1.007276) quasi with a 1 if the adduct should be considered as quasi-molecular and a 0, otherwise (Optional) polarity with a character vector indicating the polarity of the adduct or fragment. The options are "positive" or "negative"
`polarity`	Acquisition mode of the study. It can be "positive" or "negative".
`do.Par`	TRUE if parallel computing is required. Def: TRUE
`nClust`	Number of cores that may be used if do.Par = TRUE.
`IData`	Data frame containing the intensity for each sample in its columns.
`Rt`	Vector containing the retention times.
`Rt.05`	Retention time value to get a similarity of 0.5.
`use`	An optional character string giving a method for computing correlations in the presence of missing values. Default is "everything", but when missing values are present, "pairwise.complete.obs" is required.
`method`	A character string indicating which correlation coefficient is to be computed. One of "pearson" (default), "kendall", or "spearman".
`pca.to.tune`	PCA to perform the spectral clustering.
`data.prep`	Result returned by `dataPrep` function.
`k.tuned`	Optimized number of clusters (k) computed using `k.optimization` function.
`Peak.List`	Data frame containing the LC-MS features. Columns should contain: Peak.Id for a peak identifier mz for a mass-to-charge ratio value rt for the retention time Intensities for each sample
`Intensity.idx`	Numeric vector indicating the column index for the intensities
`nrow.List`	Numeric vector indicating the number of peaks.
`Annotated.Tab`	Data frame returned by `matchingStage` function.
`MH.Tab`	Data frame returned by `clusterBased.filter` function.

Function clusterBased.filter returns a data frame of filtered candidates.

Function dataPrep returns a list containing the Gaussian similarity matrices for the retention time differences and the intensities correlation.

Function eps.optimization returns an optimized epsilon parameter for dbscan algorithm.

Function featuresClustering returns the input peak list with an additional column named pcgroup that indicates the clustering.

Function k.optimization returns the ptimized number of clusters (k) using kmeans algorithm.

Function recoveringPeaks returns a data frame of filtered candidates but with all peaks recovered.

data("sample.keggDB")
Cpd.Add <- CpdaddPreparation(KeggDB = sample.keggDB, do.Par = FALSE)
data(sample.dataset)
Peak.List <- sample.dataset$Positive$Input
Annotated.List <- matchingStage(Peak.List = Peak.List, Cpd.Add = Cpd.Add,
                                polarity = "positive", do.Par = FALSE)
Intensity.idx <- seq(27,38)
clustered <- featuresClustering(Peak.List = Peak.List, 
                                Intensity.idx = Intensity.idx, 
                                do.Par = FALSE)
Annotated.Tab <- Annotated.List$Peak.Cpd
Annotated.Tab <- merge(Annotated.Tab,
                       clustered$Peak.List[,c("Peak.Id", "pcgroup")],
                       by = "Peak.Id")
                       
MH.Tab <- clusterBased.filter(df = Annotated.Tab, 
                              polarity = "positive")
                              
data(sample.dataset)
Peak.List <- sample.dataset$Positive$Input
Intensity.idx <- seq(27,38)
clustered <- featuresClustering(Peak.List = Peak.List, 
                                Intensity.idx = Intensity.idx, 
                                do.Par = FALSE)
data("sample.keggDB")
Cpd.Add <- CpdaddPreparation(KeggDB = sample.keggDB, 
do.Par = FALSE)
data(sample.dataset)
Peak.List <- sample.dataset$Positive$Input
Annotated.List <- matchingStage(Peak.List = Peak.List, Cpd.Add = Cpd.Add,
                                polarity = "positive", do.Par = FALSE)
Intensity.idx <- seq(27,38)
clustered <- featuresClustering(Peak.List = Peak.List, 
                                Intensity.idx = Intensity.idx, 
                                do.Par = FALSE)
Annotated.Tab <- Annotated.List$Peak.Cpd
Annotated.Tab <- merge(Annotated.Tab,
                       clustered$Peak.List[,c("Peak.Id", "pcgroup")],
                       by = "Peak.Id")
                       
MH.Tab <- clusterBased.filter(df = Annotated.Tab, 
                              polarity = "positive")
                              
recoveredPeaks <- recoveringPeaks(Annotated.Tab = Annotated.Tab,
                                  MH.Tab = MH.Tab)