RAMClust: A Novel Feature Clustering Method Enables Spectral-Matching-Based Annotation for Metabolomics Data

Abstract

Metabolomic data are frequently acquired using chromatographically coupled mass spectrometry (MS) platforms. For such datasets, the first step in data analysis relies on feature detection, where a feature is defined by a mass and retention time. While a feature typically is derived from a single compound, a spectrum of mass signals is more a more-accurate representation of the mass spectrometric signal for a given metabolite. Here, we report a novel feature grouping method that operates in an unsupervised manner to group signals from MS data into spectra without relying on predictability of the in-source phenomenon. We additionally address a fundamental bottleneck in metabolomics, annotation of MS level signals, by incorporating indiscriminant MS/MS (idMS/MS) data implicitly: feature detection is performed on both MS and idMS/MS data, and feature-feature relationships are determined simultaneously from the MS and idMS/MS data. This approach facilitates identification of metabolites using in-source MS and/or idMS/MS spectra from a single experiment, reduces quantitative analytical variation compared to single-feature measures, and decreases false positive annotations of unpredictable phenomenon as novel compounds. This tool is released as a freely available R package, called RAMClustR, and is sufficiently versatile to group features from any chromatographic-spectrometric platform or feature-finding software.

An untargeted metabolomics method for archived newborn dried blood spots in epidemiologic studies

“…Filtering for features that were present in at least 75% of all DBS, with a mean fold change > 3 compared to filter-paper extracts, and with a CV <25% in the pooled-QC injections, reduced the number of features to 3,157. Additional pre-processing was performed with MetMSLine software (Edmands et al 2015) to impute any remaining zero values with half the minimum non-zero feature abundance, to remove outliers by PCA, based on a tolerance of 95% beyond the Hotellings T2 ellipse, and to eliminate redundant signals due to in-source fragmentation and product formation (Broeckling et al 2014). This reduced the total number of testable features to 1,107.…”

A Key Role for Apoplastic H2O2 in Norway Spruce Phenolic Metabolism

“…Profiling of the cells and the medium yielded 968 and 551 m/z features that could be reduced further to 430 and 225 m/z peak groups, respectively. However, 75 and 53 m/z features could not be grouped and remained as singletons (Broeckling et al, 2014;Morreel et al, 2014). Statistical analyses were performed on the pseudomolecular ion from each peak group and on the singleton ions (i.e.…”

Annotation: A Computational Solution for Streamlining Metabolomics Analysis

“…Finally, this approach also needs a correlation threshold to be determined. Variations on these approach include RAMClust [28], that uses an hierarchical clustering-based approach to group both MS and MS/MS peaks, or xMSannotator [29], which uses a weighted correlation network analysis and does not require a minimum correlation threshold to be defined.…”

“…In-source fragments can be used not only to determine the monoisotopic mass where few or any adducts have been found, but also to provide with a more specific list of candidate metabolites, as each of these fragments are specific of a small number of metabolites. Examples of adoption of this strategy includes RAMClust [28], or the proof of concept by Lynn et al [33], where they matched in-source fragments with low energy MS/MS spectra in public databases. Similarly, while attempting to predict in-silico in-source fragments and retention time from molecular descriptors, the STOp-1 [34] algorithm uses in-source fragments to rank the list of putative identities.…”

“…As described before, an example of integration of MS/MS information to annotate is considering in-source fragments, which can be retrieved from low energy MS/MS spectra. For example, RAMClust [28], reduces false positive annotations by jointly processing both MS 1 and data-independent MS 2 data, whereas Scheubert et. al.…”