Everest Manual
User Manual:
Open the PDF directly: View PDF 
.
Page Count: 5
An introduction to Everest package
Version 0.3.0
Xavier Domingo-Almenara (Maintainer)
xdomingo@scripps.edu
May 23, 2018
This vignette presents Everest, an R package for annotation of metabolites in untargeted
LC–MS-based metabolomics. Everest allows annotation of a list of features from XCMS or other
computational tools. Everest provides a full characterization of features, detecting adducts,
common losses and dimmers, to computationally determine monoisotopic masses from underlying
metabolites in the samples. Everest provides a list of putative identification based on accurate
mass search of computed monoisotopic masses. If you use the package Everest in your analysis
and publications please cite:
Domingo-Almenara X. Everest, an R package for metabolite annotation in untargeted LC-MS-
based metabolomics. CRAN. URL https://CRAN.R-project.org/package=everest
Installation: Everest can be installed from any CRAN repository, by:
# Install
install.packages("everest")
# Load
library(everest)
Support: Any enquiries, bug reports or suggestions are welcome and they should be addressed
to xdomingo@scripps.edu.
1
1 Introduction
Everest is an R package that allows the annotation of LC–MS-based untargeted data. Everest
will first group those features that stem from the same metabolite, assigning to each one an
’AlignID’ number. All these features sharing the same ’AlignID’ number are considered a single
pseudo-spectra resulting from the same molecule. Next, Everest will annotate these features
within each pseudo-spectra, using known monoisotopic masses for common adducts, losses and
dimmers. Features that show a m/z relation will be annotated with the same ’AnnID’ number.
2 Importing feature tables
Everest can work with xcmsSet object directly from XCMS, or import tables. If you have pro-
cessed your data with XCMS (Offline version), and you have the xcmsSet (e.g., xset3) loaded into
your R enviroment, then proceed to the next section. Otherwise, if you have processed your data
and you have a CSV or Excel file (e.g., diffreport from XCMS Online) then read the following
steps.
First, edit the spreadsheet (Excel or CSV) so that the first column contains the features names,
the second the m/z values and the third the Retention Times, and name them as: name, mz and
RT, respectively. The rest of the columns should contain only the intensity or area (recommended)
values of each feature in each sample. Remove any other columns or rows. Here is an example of
how the table should look like:
name mz RT Sample1 Sample2 ...
M50T350 50.22695 349.5595 375013.1 1997693.4
M51T316 51.28015 316.2429 2278437.1 213624.8
M51T349 51.28734 349.3183 23420.0 3988480.2
M51T3162 51.30070 316.3434 1800327.7 2046424.4
M51T3163 51.31288 316.4545 1208414.6 1546696.9
M53T350 52.94600 349.5218 318827.2 1227462.8
+RT and column names: The first column and the sample column can have any
name. However, columns 2 and 3 have to follow this exact spelling: mz and RT. Also,
the RT must be in seconds, so multiply the RT (rtmed) column per 60 if the data is
in seconds (specially when using XCMS Online, as it is the default option)
Next, the file can be loaded into R using the native read.csv2() or the read.xlsx() function
from openxlsx package1.
# Load table (CSV)
feature.table <- read.csv2("featureTable.csv")
# Load table (tab delim TSV file)
feature.table <- read.delim("featureTable.tsv")
# Or an XLSX file:
feature.table <- read.xlsx("featureTable.xlsx")
# Make sure that it looks OK by taking a quick look:
head(feature.table)
If the output of head(feature.table)looks good, then we can proceed to the following
section. Be careful, sometimes when loading the table, (e.g., if using TSV files) an additional
column is introduced. We can remove it easily with R (see following section)
1You can install openxslx (https://CRAN.R-project.org/package=openxlsx ) by typing the following into the R
console: install.packages(”openxlsx”)
3
3 Annotation of features in untargeted LC/MS-based data
In this section we show the annotation of a feature list as a result of the analysis of blood samples
in positive mode with processing via XCMS. The feature table used to reproduce this example
can be download from the Github repository 2. However, the following code includes a script to
automatically download the file. All the listed commands (script) to reproduce the following demo
can by found by executing:
help(package="everest")
and then click on User guides, package vignettes and other documentation and on source from
the ’Everest Manual’.
3.1 Metabolite annotation
The following code will download the DEMO feature table and proceed with the metabolite an-
notation.
library(everest)
# Download the feature table (if not downloaded manually before):
download.file("https://github.com/xdomingoal/everest-data/raw/master/MTBLS20.tsv"
, "MTBLS20.tsv")
# Load the table (we have to remove the first column of the table, as explained be-
fore):
feature.table <- read.delim("featureTable.tsv")[,-1]
# Annotate
ex <- evAnnotate(data.table = feature.table, ion.mode ="pos",
min.correlation = 0.7, max.time.dist = 2, ppm.error = 20)
The results can be accessed throught the function annoTable(), and we can save the resulting
file as an Excel or CSV file:
# Access the annotation results:
anTab <- annoTable(ex)
# Write the table as an excel file:
require(openxlsx)
write.xlsx(anTab, file ="results.xlsx")
The resulting table will contain four additional columns:
AlignID: All the features sharing the same AlignID are considered to stem from the same
metabolite.
AnnID: Within the same AlignID, these features sharing the same AnnID are considered to
stem from a unique monoisotopic mass. A unique AlignID may have more than one AnnID,
and each AnnID number leads to a different monoisotopic m/z (see example below).
Isotope: Indicates whether or not this feature is an isotopic peak.
toMSMS: Indicates whether or not Everest recommends these feature for subsequent tandem
MS fragmentation.
2URL to download the file. https://github.com/xdomingoal/everest-data/raw/master/MTBLS20.tsv (click on
File/Save File As on your web explorer if the file does not automatically downloads.)
4
To annotate a xset3 object from XCMS (offline version), use the following code:
# Annotate an xset3 object from XCMS
ex <- evAnnotate(xcmsSet = xset3, ion.mode ="pos",
min.correlation = 0.7, max.time.dist = 2, ppm.error = 20)
3.2 Ohter functionalities
Two other functions are included in , showAn and showGroup. The first displays access to a
summary of the annotation results for a specific feature. Let us consider that we want to know
the annotation of the feature ’M170T276 8’, in that case:
showAn(ex,"M170T276_8")
AlignID AnnID Feature Mass Rt Adduct IsoCount Score mIsoMass Isotope toMSMS
525 1 M152T276 152.0344 276.4 (M+H)+[-H2O] 1 99 169.0382
525 1 M170T276_8 170.0448 276.4 (M+H)+ 0 99 169.0382 yes
525 1 M192T276 192.0268 276.1 (M+Na)+ 0 99 169.0382
525 1 M339T276 339.0852 276.1 (2M+H)+ 0 99 169.0382
525 1 M153T276 153.0376 276.3 M+1 0 99 169.0382 yes
525 2 M152T276 152.0344 276.4 (M+Na)+[-H2O] 1 99 147.0562
525 2 M170T276_8 170.0448 276.4 (M+Na)+ 0 99 147.0562
525 2 M192T276 192.0268 276.1 [M-H+2Na]+ 0 99 147.0562
525 2 M153T276 153.0376 276.3 M+1 0 99 147.0562 yes
In this example, the queried feature is shown together with the other features sharing the same
AlignID number. We observe that the queried feature is a protonated (M+H) specie, and that
shares its AnnID number with other four features, which correspond to adducts, losses, dimmers and
isotopes. Additionally, other four features within the same AlignID group share a different AnnID
number. This indicates that these AlignID group could actually be another metabolite (protonated
specie). Everest ranks the different annotation hypothesis, and in this case, proposes the queried
feature for subsquent tandem MS fragmentation (toMSMS column). The Isocount column indicates
how many isotopic peaks have been found for that feature, whereas the mIsoMass indicates the
monoisotopic mass to which all the features whithin a unique AnnID point to. Of note, decimals
in this example are truncated for illustration purposes.
The second function, showGroup retrieves a summary of the group (AlignID) of that feature,
and it contains specific information such as the pseudo-spectra or the list of features that compose
that group.
5
