Datasets

The metabodecon repository contains a selection of example datasets. This article describes each of these datasets in details, i.e.

• which and how many samples are included
• how they were measured
• how you can access the dataset

The Blood dataset

The blood dataset contains 16 one-dimensional CPMG NMR-spectra of human blood plasma in Bruker format. It can be found in folder misc/datasets/blood in the metabodecon repository.

The Urine dataset

The urine dataset contains two one-dimensional NOESY NMR-spectra of urine, available in both Bruker and jcamp-dx format. They can be found in folder misc/datasets/urine in the metabodecon repository.

The Sim dataset

There are scenarios where it is useful to work with simulated datasets instead of real data, such as:

• When you need to know the underlying distribution of the data to check whether a function works as expected.
• To speed up test cases and examples where a few data points are sufficient to test a function.

For such cases, metabodecon includes a simulated dataset called sim, which was generated by applying the following steps to each spectrum of the blood dataset:

1. Deconvolute spectrum using generate_lorentz_curves() with default parameters
2. Extract Lorentz curve parameters for all peaks between 3.52 and 3.37 ppm
3. Generate 2048 equidistant chemical shift values between 3.59 and 3.28 ppm¹
4. Calculate the signal intensity at each chemical shift as superposition of Lorentz curves
5. Add random noise to the simulated spectrum ²

The first two of the 16 simulated spectra are plotted below. For further details about the simulation process, see the source code of function simulate_spectrum().

Figure: The first two simulated datasets from the sim.

How to download datasets

Due to the size constraints for R packages, most of the above mentioned datasets are not included by default when the package is installed, but must be explicitly downloaded afterwards. This can be done via command download_example_datasets():

library(metabodecon)
# Set persistent = TRUE to store the files at a persistent location. This way,
# the next time you call `download_example_datasets()`, the files will not be
# downloaded again.
path <- download_example_datasets(persistent = FALSE)
tree(path)

Spectra that come pre-installed with the package and do not require a separate download, are:

• All 16 spectra from the sim dataset
• The two spectra from the urine dataset in Bruker format
• The first spectrum from the urine dataset in jcamp-dx format