Systems Biology (Methods in Molecular Biology)

4 Notes

1. If you cannot prepare the fecal samples immediately after col-
   lection, store them at 80 C.


2. For each experiment, the magnetic field homogeneity must be
   optimized through an accurate shimming. To check if a sample
   is properly shimmed, you can observe the full-width at half-
   maximum of lactate peak that should be less than 1.7 Hz,
   before applying apodization and symmetric shape. TSP peak
   is not a reliable signal to check the quality of shimming due to
   the huge amount of proteins present in the sample, which
   influence its line-width because TSP binds to proteins.


3. Samples must be acquired in randomized order.


4. It is useful to run a standard solution to identify potential
   impurities arising from reagents and preparation procedures.


5. The most common method of spectral binning is the so-called
   equidistant binning, i.e., each spectrum is divided into bins
   with fixed width, typically 0.04 ppm. The weakness of this
   method is that, under certain experimental conditions, single
   peaks can be divided into two neighboring bins, generating
   artifacts. To avoid this problem, several mathematical algo-
   rithms [14–17] have been developed to vary the individual
   size bin. For example, ACD intelligent bucketing method
   (ACD/NMR processor, Advanced Chemistry Development
   Inc. (ACD/Labs), Toronto, ON, Canada), a combination of
   equidistant binning and non-equidistant binning, sets the
   bucket divisions at local minima (within the spectra) to ensure
   that each resonance is in the same bin throughout all spectra.

Table 4
Pearson’s correlation analysis among PC1, PC4, clinical status (healthy or sick), and pancreatic
insufficiency factor (PI)

PC1 PC4

Pancreatic

insufficiency factor

Clinical status (healthy

or sick)

PC1 1.000 0.012

(0.334)

0.388 0.382 (0.120)

PC4 0.012

(0.334)

1.000 0.604 0.592 (0.210)

Pancreatic

insufficiency factor

0.388 0.603 1.000 0.822

Clinical status (healthy

or sick)

0.382

(0.120)

0.592

(0.210)

0.822 1.000

In parenthesis, partial correlation coefficients among PC1, PC4, and clinical status adjusted for the PI effect are reported

Metabolomics and Clinical Needs 335