Biology Reference
In-Depth Information
large-scale GC/MS-based studies clearly require
extensive validation to ensure the high-standard
quality of data and validity of results. The
following section describes speci
performance, both sample preparation and anal-
ysis steps should be subjected to randomization
so as to reduce bias associated with the
processes.
19,20
Because instrument contamination, column
degradation, and detector aging are factors
known to affect reproducibility of data and
contribute to analytical drift, it is recommended
to perform instrument maintenance at the start
of each analytical block. For example, consum-
ables such as syringe, injector liner, septum,
and gold seal may be replaced; the tip of GC
column (5 cm) may be cut to reduce cross-talk;
and tuning and mass calibration of GC/MS are
performed.
19,47
c strategies
adopted in our lab to address large-scale GC/
MS-based metabonomic studies.
Methodological Considerations in Sample
Preparation and Analysis
Analytical drift in both chromatographic and
mass spectrometer performance is well reported
in the literature. This challenge is accentuated in
large batch analysis, leading to a compromise in
the instrument reproducibility.
19,44
e
46
Such
analytical drift has been reported to be due to
the analytical cross-talks, changes in tempera-
ture, and instability of electrical circuitry.
In large-scale studies, not all samples could be
analyzed in a single analytical batch due to co-
sharing of instrument and downtime related to
maintenance. As such, samples are categorized
into multiple analytical blocks to be analyzed
using a single instrument or multiple instruments.
The sample size in each blockmust be be validated
to ensure reproducible data within a stipulated
period of analysis. The strategy of performing
subsets of analytical experiments to obtain repro-
ducible data and then integrating the data from
multiple analytical experiments into a single data
set requires the analysis of QC samples
throughout the entire studyduration.
19,47
Indeter-
mining the block size, apart from the stability of
the analytical system, analysts should also
consider the number of samples that can be
prepared per day, analytical run time, instrument
time, stability of samples, and schedule for
instrument maintenance. Begley et al. validated
the use of GC/TOFMS in the analysis of human
serum samples for long-term metabolomic stud-
ies; 120 clinical samples (and associated 60 QC
and 8 blank samples) were deemed appropriate
for a single analytical block (GC/MS run time of
25 min per sample).
19,47
To eliminate systematic
bias due to the gradual change in instrument
Quality Control
Due to the extensive sample preparation and
long analysis time associated with large-scale
studies, QCs are incorporated to monitor the
performance of the method.
4,48
Several strate-
gies have been proposed, including the use of
external standards (pooled QCs or standards
without matrix), internal standards, or a combi-
nation of both internal and external standards.
49
Although standards without matrix can be
used to detect the decline in system perfor-
mance, pooled QCs remain more commonly
adopted. Pooled QCs are identical biological
samples obtained by pooling those biological
samples under study or commercially available
bio
uids not present in the study.
19,48
The
former is recommended, as the metabolic
composition of commercial bio
uids is not
representative of the study population. In
large-scale studies in which sample preparation
and analysis commence before sample collection
is completed, QC samples may be pooled from
a subset of the study population. Pooled QCs
are subjected to the entire metabonomics work-
flow from sample preparation to data analysis
and serve several functions.
4,9,19,20,47,48,50
First,
pooled QCs are used to condition the GC/MS
system before actual samples are analyzed to
Search WWH ::
Custom Search