ramClustMod: 'ramclustR' - modified for MetMSLine

In WMBEdmands/MetMSLine: an automated and fully integrated pipeline for rapid processing of high-resolution LC-MS metabolomic datasets.

Description

ramclustR - modified for MetMSLine

Usage

ramClustMod(peakTable = NULL, obsNames = NULL, st = 3, sr = 0.5,
  maxt = 20, mult = 5, deepSplit = FALSE, blocksize = 2000,
  hmax = 0.3, minModuleSize = 2, linkage = "average")

Arguments

peakTable	optional if pcaResult argument not supplied. Either a data.frame, full file path as a character string to a .csv file of a peak table in the form observation (samples) in columns and variables (Mass spectral signals) in rows. If argument is not supplied a GUI file selection window will open and a .csv file can be selected.
obsNames	character vector of observation (i.e. sample/ QC/ Blank) names to identify appropriate observation (sample) columns.
st	Sigma t value controlling the rate at which feature similarity decays with respect to differences in retention times between feature pairs. When an xcms diff report is used as input, this parameter is by default set to half the peak width of all the peaks in the xcmsSet. Can be manually overridden to tune performance when necessary. Larger values are less selective, smaller values are more selective.
sr	Sigma r value controlling the rate at which feature similarity decays with respect to correlational simiarity between a pair of features. numeric e.g. 0.5, Sigma r: the rate of similarity decay in response to correlational r . Can be manually overridden to tune performance. Larger values are less selective, smaller values are more selective.
maxt	numeric e.g 20, the maximum time allowed for similarity between features to be greater than zero. For example, if two features are separated by more than 20 seconds, the similarity score is automatically set to zero, ensuring that they end up in different clusters. If you are importing data from another vendor, ensure that either you change the feature times to seconds (default in ramclustR) or you change this value to a value representing minutes, rather than seconds (e.g. 0.33). Used primarily to speed up compuation time by not calculating values above this threshold.
deepSplit	boolean e.g. =FALSE, access to the deepSplit function in dynamicTreeCut. False generates fewer spectra, TRUE splits spectra more readily and thereby generates more spectra. For more detailed description, see documentation for the dynamicTreeCut package.
blocksize	integer number of features processed in one block =2000, ramclustR makes attempts to minimize memory demands to maintain processing speed by avoiding memory constraints. It does so by processing the full similarity matrix in blocks of blocksize dimensions. This has two benefits: one is that a given PC is less likely to run out of memory during processing and the second is that we need not process blocks that contain not feature pairs with a delta rt of greater than maxt, thereby avoiding needless calculations and processing time.
minModuleSize	the number of features required for a cluster to be reported. That is, if a feature is clustered in a group with less than minModuleSize features, it will not be exported. If you wish to analyze singletons, set minModuleSize=1.
linkage	The linkage method used to perform heirarchical fastcluster-based heirarchical clustering. default="average".

Value

a list containing 2 elements:

1. the original peak table with the results of retention time correlation clustering included in additional column "RtCorrClust" as a data frame.

2. A data frame of the weighted means and details of the most intense feature for each group. In addition for each group details of the best sample to consider for subsequent MSn fragmentation analysis based on the most top 5 most intense samples is provided. This "best" sample for MS/MS is also selected whilst concurrently minimising the number of samples to potentially reinject for targetted MSn fragmentation analysis.

Source

http://pubs.acs.org/doi/abs/10.1021/ac501530d RAMClust: A Novel Feature Clustering Method Enables Spectral-Matching-Based Annotation for Metabolomics Data.

WMBEdmands/MetMSLine documentation built on May 9, 2019, 10:03 p.m.