Biology Reference
In-Depth Information
assess system suitability and ensure reproduc-
ible results. Five
principal components indicates analytical drift
in signal, retention time shifts, or sample degra-
dation or evaporation. To overcome the
problem, those peaks responsible for the drift
could be identi
injections of QCs are
recommended prior to GC/MS analysis to
equilibrate the active sites of the system with
sample matrix especially postpreventive main-
tenance.
19,47
Second, pooled QCs serve as
a measure of repeatability and are used to
monitor the analytical variation. QCs are
analyzed at the beginning, interspersed at reg-
ular intervals between study samples within an
analytical block and at the end of the sample
queue. The raw chromatographic data of QCs
are visually inspected to detect drift in peak
intensities and retention time in which selected
peaks could be examined for parameters such
as peak shape, signal intensity, and retention
time to detect for gross changes in the system.
If a system problem is detected, the necessary
troubleshooting is performed and the affected
samples are reanalyzed. A tolerance of 30% CV
in technical precision of each metabolic feature
in the QCs has been the acceptable variation
considering the untargeted nature of the analyt-
ical method in detecting metabolites in the bio-
logical matrices.
19,47,51
A data set containing
more than 80% of metabolic features with CV
less than 30% would be considered good
quality.
50
As pooled QCs are identical biological
samples, peaks in QCs with poor repeatability
of more than 30% CV are removed before data
analysis. If the method is validated, QCs are
expected to cluster closely in the PCA score
plot. One may ask how tightly clustered the
QCs should be for the method to be acceptable.
The current practice is to use QCs to reject
analytical batches, in which highly variable or
scattered QCs in the PCA score plot would indi-
cate analytical failure.
48
For metabolites identi-
“
lead-in
”
ed by constructing a new model
based only on the QCs and the drift may be cor-
rected via re
nement of peak alignment, data
normalization, or exclusion of nonreproducible
peaks.
52
Hence, the third function of pooled
QCs is to provide data for signal correction
within and between analytical blocks.
19
Fourth,
pooled QCs function as reference by which the
metabolic features are used in the alignment of
the metabolic features of other samples.
20
Manual exclusion of interfering peaks of drugs
or their metabolites from the metabolic pro
les
is tedious and erroneous. In such cases, a useful
strategy that our group has adopted is to pool
QCs solely from control samples in which no
drug treatment was administered. This approach
ensures complete exclusion of all drug-related
metabolic features so that any differentiation in
the metabolic pro
les of the control and treat-
ment group is not due to the occurrence of
drug or its metabolites but rather the biological
effects. However, a limitation of this approach
is that the method may exclude metabolites
that are highly elevated in the treatment groups
yet below the detection limit in the controls.
For the internal standard method, exogenous
or isotopically labeled standards are spiked into
every sample. An exogenous standard (a stable
analyte not derivatized and not present in the bio-
logical matrices) helps monitor injection volume,
detector sensitivity, and correct detector
response.
49
The addition of labeled metabolites
prior to storage, extraction, derivatization, or
analysis enables the control of these steps in
sample workup.
49
An endogenous metabolite in
the sample can be corrected using the isotopically
labeled form of the metabolite or another metab-
olite of similar reactivity towards silyation and
stability postderivatization. However, in untar-
geted metabonomics, it is not viable to prepare
a complete mixture of internal standards due to
fied as biomarkers, analysts are advised to
examine their variability in the QC data and
the analytical variability should be less than the
effect by treatment or disease. The spread of
QCs in the PCA score plot can reveal issues
such as time-related drift. Observation of injec-
tion order of the QCs in the
rst and second
Search WWH ::
Custom Search