Biomedical Engineering Reference
In-Depth Information
metabolomics represents a 'snapshot' at one point in time and gives no
information about the dynamic fl ux of metabolites. For example, a
metabolite showing a small concentration may in fact be transported in
large quantities or one that shows high quantity may be produced very
slowly but accumulates over time.
Identifi cation of metabolites from the mass spectrum is one of the
biggest challenges in metabolomics. Accurate mass and fragmentation
patterns may assist in the determination of structure but for absolute
certainty, preparative isolation of the substance followed by other
methods such as NMR may need to be employed. Therefore in many
studies it must be realised that identifi cations are somewhat tentative.
This is not necessarily a major problem in hypothesis-generating research,
but could potentially limit the use of untargeted methods in critical areas
such as clinical diagnosis. Fortunately, robust multivariate methods may
be used to fi ngerprint the combination of many metabolite signals in
order to produce classifi cations with high levels of accuracy (low false
positives and negatives) so the precise identifi cation of individual
components may be unnecessary for some applications.
4.4 Data types
Occasionally mass spectrometry data may be in the form of a mixed
spectrum of all components in the case of 'direct infusion' where a sample
is infused into an instrument. This is a simple 2D mass spectrum of mass/
charge ratio (m/z) and ion abundance (intensity).
More commonly data come from LC-MS, GC-MS or CE-MS where the
complex mixtures are separated before being introduced to the mass
spectrometer. The data which result are three-dimensional, having axes of
mass/charge ratio (m/z), retention time and ion abundance (intensity).
4.4.1 Vendor formats versus open formats
The data from mass spectrometry, in common with many areas of analytical
chemistry, are often saved in proprietary formats dictated by instrument
vendors. There are several disadvantages with this situation, the diffi culties
of using a common software platform in labs with different manufacturers'
equipment, the hindrance of free interchange of data in the scientifi c
community and the lack of ability to read archived data in the future.
Although it is unlikely that instrument vendors will accept a common fi le
Search WWH ::
Custom Search