SpikedApple: Spike-in metabolomics data for apple extracts
In BioMark: Find Biomarkers in Two-Class Discrimination Problems

Description Usage Format Details Author(s) Source See Also Examples

Data from a spike-in experiment for apple extracts. Twenty apple extracts are divided in two groups, one control, and one spike-in group. The control group is measured without any spiking - the spike-in group is spiked with nine chemical compounds in three different combinations of concentrations. The data provide the experimental data of the forty apple extracts in lists SpikePos and SpikeNeg for positive and negative ionization, respectively, and in two separate data.frames (pos.markers and neg.markers) contains information of the features of the standards, i.e., the spike-in compounds.

1 2	data(SpikePos) data(SpikeNeg)

SpikePos and SpikeNeg are lists with three elements:

data: Data matrix, describing for each of the forty injections the intensity of the features (columns). Column names consist of a combination of retention time (in seconds) and m/z values, and are sorted on retention time.
classes: Class labels for the forty injections (control, or group1, 2 or 3).
annotation: Matrix, containing for each of the features XCMS and CAMERA information, such as mz, rt, number of times a feature is identified in the control or spike-in samples, possible isotope or adduct annotation, and whether or not the feature is identified in the standards (the spike-in data).

In addition, pos.markers and neg.markers contain the information of the standards, i.e. the compounds that are spiked in. These data.frames describe in their rows single features identified with XCMS and CAMERA, using the same settings as the experimental apple data, and have the following columns:

comp: The (short) name of the spiked-in compound giving rise to this particular feature.
mz, rt, isotope, adduct: Feature information, similar to the information in the annotation fields in SpikePos and SpikeNeg.
feature.nr: The number of the corresponding feature in either SpikePos or SpikeNeg.
group1, group2, group3: Approximate spiking levels for the three groups. A value of 1.0 corresponds to an increase that is roughly equal to the naturally occuring concentration in apple. Exceptions are trans-resveratrol and cyanidin-3-galactoside, both not naturally occuring. These two compounds have been spiked in at one constant level which gives features of comparable size.

This is the complete data set, from which spikedApples is a subset, basically presenting the control and group1 information with hand-picked spike-in features. The data in SpikePos and SpikeNeg use CAMERA grouping to automatically determine which features are corresponding to which spike-in compounds. Raw data in CDF format are available from the MetaboLights repository.

Pietro Franceschi

http://www.ebi.ac.uk/metabolights/MTBLS59

P. Franceschi, D. Masuero, U. Vrhovsek, F. Mattivi and R. Wehrens: A benchmark spike-in data set for biomarker identification in metabolomics. J. Chemom. 26, 16-24 (2012).

spikedApples

data(SpikePos)
plot(SpikePos$annotation[,c("rt", "mz")],
     xlab = "Time (s)", ylab = "m/z",
     main = "Positive ionization mode")
points(pos.markers[!is.na(pos.markers$feature.nr), c("rt", "mz")],
       pch = 19, col = 2)

data(SpikeNeg)
plot(SpikeNeg$annotation[,c("rt", "mz")],
     xlab = "Time (s)", ylab = "m/z",
     main = "Negative ionization mode")
points(neg.markers[!is.na(neg.markers$feature.nr), c("rt", "mz")],
       pch = 19, col = 2)