Researchers have created a public repository containing nearly 24 million tandem mass spectra from a single molecule to show why the number of spectra alone cannot describe the chemical coverage of a mass spectrometry database. The resource, called GinorMass, contains 23,971,300 spectra of arginine, an amino acid involved in several biological processes.
Mass spectral libraries are widely used to help identify molecules in biological, clinical, environmental, and chemical samples. An unknown spectrum can be compared with reference data to find a possible match. The value of this approach depends partly on the range of authenticated compounds represented in the library.
Repositories often use their total number of spectra, files, or datasets to communicate their size. However, millions of spectra can include repeated measurements of the same molecules. A high spectral count therefore does not necessarily mean that a resource covers a wide range of compounds.
GinorMass demonstrates this difference directly. Although it contains almost 24 million experimentally acquired spectra, they all represent one authenticated molecule. A database with the same number of spectra collected from many verified compounds would contain much greater chemical diversity.
The researchers used a 10-µg/mL arginine reference standard and a quadrupole time-of-flight mass spectrometer. The sample was introduced by syringe pump at a flow rate of 3 µL/min. Spectra were collected in positive-ion mode at a collision energy of 20 eV, with each acquisition taking between 20 and 50 milliseconds.
Arginine was selected because it is stable, readily available, biologically relevant, and well characterized. Under the chosen conditions, it produced five fragment ions, placing it within the typical fragmentation range of compounds in a large reference library.
Repeated measurements can still be valuable. Spectra collected with different instruments, laboratories, collision energies, and sample conditions can reveal analytical variation and help laboratories assess whether results are reproducible and transferable. Under one fixed set of conditions, however, further measurements may become redundant.
The authors recommended that repositories report spectral count alongside measures that describe their scientific content. These include the number of unique molecules, the number confirmed with reference standards, the chemical classes represented, instrument and collision-energy coverage, ionization conditions, annotation confidence, and data provenance.
Each spectrum should also be linked to its sample source, acquisition conditions, and processing history. This information could help users judge whether a reference is suitable for a particular analytical method or identification task.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
