SUPER SESI generates data about the molecular composition of breath.
The ultimate characteristic of an analyzer is the quality of the data that it generates.
what is High quality data?
Ideally, a data set should reflect the original composition of the analyzed sample, down to its finest detail.
Limits of Detection. The LoD are important to detect species with very low concentration. For this, SUPER SESI provides
Identification: The separation capacity is required to identify the different compounds and determine their chemical formula. The fact that SESI is non fragmenting greatly simplifies the spectra and its interpretation. SUPER SESI is tailored for high resolution mass spectrometers, to enable for the determination of the chemical formula directly from the molecular mass data.
Time resolution: The time resolution of the ionizer and the analyzer helps to identify causality correlations.
However, signals other than the analyzed sample also contribute to the data. These signals may include:
Background signals produced by the instrument. Usually, these signals are well known and characterized precisely because they are constantly produced. Background signals deteriorate the LoD, but they can be easily differentiated from sample signals and eliminated by most data post-processing algorithms.
Other confounding variables. Confounding variables are signals that have the same appearance as the sample of interest, but do not belong to it.
A high quality data set provides a fine description of the analyzed samples, has a low background signal, and it is free of confounding variables.
why is data quality important for breath analysis?
Discovering breath biomarkers has proven to be a difficult task, with many researchers not being able to replicate their results.
Confounding variables lead to spurious correlations. To avoid this, the data should reflect the analyzed samples and be free of other confounding variables.
Confounding variables appear as bio-markers, but they are not.
Ibuprofen is a usual suspect, and a good example. Higher levels of ibuprofen correlate with disease, but this doesn't mean that it is a reliable biomarker.
Less obvious confounding variables might be lurking in other steps of a study: For instance, different handling of patient and control samples might jeopardize an entire clinical study.
with increased sensitivity, High quality data is a must
Things get increasingly complicated with high sensitivity instruments because they detect a larger number of species.
Supervising very large data-sets with large number of species becomes impossible for a human.
As a result, automated and more sophisticated data-mining tools are required.
But integrating the knowledge required to differentiate confounding variables from legitimate biomarkers is a challenge.
Due to the improved Limits of Detection, the probability of inadvertently introducing a confounding variable while handling the samples increases dramatically.
In practice, the quality of the data is known once the data is analyzed in detail.
Confounding variables are only identified when the study is well advanced and it can be too late to correct the sample acquisition protocols.
Traditional handling steps are a source of confounding variables that could potentially kill the efforts put in designing a biomarker discovery campaign.
Cutting the number of steps required to analyze a sample eliminates the associated risks.
SUPER SESI minimizes confounding variables
For the last two years, our efforts have been dedicated to improving the quality of the breath data. Our view is that this preliminary step is crucial to the discovery of forthcoming breath biomarkers.
This is achieved by reducing background and memory effects, and by simplifying the data acquisition protocol.
Breath is analyzed in real time, right as it is exhaled.
No sample preparation
No column pre-seaparation
The probability of introducing a confounding variable in any of these steps is eliminated.
- SUPER SESI data only reflects the sample and the instrument background.
- The instrument background is constantly measured between sample aqusitions.