Metabonomics SOP

 

Safety

Everyone working in the NMR laboratory in ‘Statler Commissary’ is expected to act professionally and environmentally responsible. Our laboratory is committed to minimize the risk of personal and environmental hazards involved in working with high-field magnets and handling biological fluids. Thus, the protocols herein are established. These protocols are specifically tailored to our laboratory and subject to change if better procedures are identified in minimizing such risks.

Contact information

---

Dr. Khalid Ahsan

• Chemistry Department Safety Officer
• ahsan@buffalo.edu
• 716-645-4115

Dr. Leonard Borzynski

• Manager, Biosafety Division
• Environment, Health & Safety (EH&S)
• lenb@buffalo.edu
• 716-829-3301; 716-829-2401 (sharps injury reporting)

University Police

• 716-645-2222

Mr. Ryan Sajdak

• PhD student
• rasajdak@buffalo.edu
• 716-474-1916

Ms. Victoire-Grace Karambizi

• MS student
• victoire@buffalo.edu
• 716-507-3880

Dr. Dinesh Sukumaran

• Director, NMR Center
• dks@buffalo.edu
• 716-725-1107

Dr. Thomas Szyperski

• Principal Investigator
• szypersk@buffalo.edu
• 716-472-7075

Szyperski Lab

• 716-645-4302, 716-645-4303, 716-645-4304, 716-645-4312

University at Buffalo Health Services

• 716-829-3316

Millard Fillmore Suburban Hospital

• 1540 Maple Road, Williamsville, NY 14221
• 716- 568-3600

STERICYCLE

• http://www.stericycle.com/medical-waste-disposal/customer-pickup-request.html
• 1-866-783-7422

NMR Safety Rules

---

Cryogenic liquids

1. No person is allowed to use cryogenic liquids without training and permissio
2. Dr. Sukumaran should be notified first before any attempt to use cryogenic liquids
3. Protective glasses and thermal gloves should be used when working with cryogenic liquids
4. In the event of a magnet quench, all occupants must exit the NMR lab immediately
5. Any accident of exposure to cryogenic liquids must be reported to Dr. Sukumaran and Dr. Khalid Ahsan.
   

Magnetic materials

1. Persons with metallic implants (e.g. cardiac pacemakers, orthopedic pins, plates, arterial clips, etc.) must keep at least 10 feet distance from the magnets
2. Metal and magnetic objects (e.g spatulas, carts, keys, ATM cards, wristwatches, laptops, magnetic tapes, disks, etc.) must be left in the wet lab room or on the table outside the NMR room
   

Physical contact

1. Don’t lean on the magnet
2. Use the stairs provided for the spectrometers when loading the sample
   

Injury reporting

1. Report any injury to Dr. Sukumaran and Dr. Ahsan

General Rules

1. In the event that the sample was broken inside the probe, immediately inform Dr. Sukumaran. Isopropyl alcohol must be used as a cleaner and disinfectant for the probe. Don’ t use bleach! Probe cleaning should only be performed with Dr. Sukumaran. If the sample is serum and if the probe/spectrometer needs fixing, inform the person who will fix the problem the potentially hazardous nature of the sample before he starts to work.
2. Use personal protective equipment (e.g. double purple nitrile gloves, lab coat, safety glasses).
3. Prepare 10% (v/v) bleach-water solution before handling serum samples. Keep the solution only for a week.
4. For NMR tube breakage, recover as much sample as possible from the tube, disinfect the broken tube by soaking it in freshly prepared bleach solution and disinfect the surface. Throw the disinfected glass fragments in the broken glass container.
   
5. In the event of blood serum spill, cover the spill with freshly prepared bleach solution and let it stand for at least an hour before wiping it off with paper towels. Redisinfect the surface with freshly prepared bleach solution. Throw the used paper towels in the biohazardous waste container.
6. If you cut or puncture yourself while handling potentially biohazardous biofluids, wash the ezposed area with soap and water and seek medical attention from Milliard Filmore Suburban Hospital or from Health Services as soon as possible (refer to Safety Protocols for NMR for injury logs and reporting).

Biosafety

---

Safety training

Schedule for a discussion on the following should be made with Mr. Garcia (et. al.)

1. Exposure Control Plan
2. Metabonomics lab set up in Statler
3. Occupational Safety and Health Administration standards Regulations (Standards - 29 CFR) Bloodborne Pathogens
4. Questions regarding other aspects of biosafety can be directed to Mr. Leonard Borzynski of EH&S.

Training records

Write a paragraph on the WIKI regarding the date, location, resource person and bullet points of biosafety topics discussed

Vaccination

There are three shots that need to be legally administered to people who are front liners (have direct physical contact with blood and blood products):

1. Undergrad/Graduate students: Call University at Buffalo Health Services to schedule an appointment.
2. Post-docs: Make sure that you have an active health insurance. Search online for health providers that are affiliated with your current health insurance and schedule an appointment for vaccination. If you have a family doctor, he is also allowed to administer the shots, if he agrees.
   

Universal safety precautions and lab safety policies

1. Prepare 10% (v/v) bleach-water solution before handling serum samples. Keep the solution only for a week. Fill out the form that is posted beside the metabonomics hood for bleach solution preparation indicating the date and name of person who prepared it
2. Wash hands with hand soap and water before and after sample preparation
3. Wear double purple nitrile gloves whenever you are working with blood and other potentially infectious materials
4. Wear full body lab coats at the time
5. Wear safety glasses
6. Dispose used pipet tips, Eppendorf tubes and gloves that came in contact with serum and serum products in their designated biohazardous waste container. Contact STERICYCLE to schedule for pick up if these containers get filled. Note that the schedules should be made at least 2 days prior to pick up dates.
7. Disinfect contaminated safety glasses, lab coats and clothing that were contaminated with serum and serum products by soaking them in a freshly prepared 10% (v/v) bleach-water solution
8. Eating and drinking is strictly prohibited inside the NMR lab
9. In the event of blood serum spill, cover the spill with freshly prepared bleach solution and let it stand for at least an hour before wiping it off with paper towels. Redisinfect the surface with freshly prepared bleach solution. Throw the used paper towels in the biohardous waste container.
10. For NMR tube breakage, recover as much sample as possible from the tube, disinfect the broken tube by soaking it in freshly prepared bleach solution and disinfect the surface. Throw the disinfected glass fragments in the broken glass container
11. In the event that the serum sample was broken inside the probe, immediately inform Dr. Sukumaran. Isopropyl alcohol must be used as a cleaner and disinfectant for the probe. Don’ t use bleach. Probe cleaning should only be performed with Dr. Sukumaran. If the probe/spectrometer needs fixing, inform the person who will fix the problem the potentially hazardous nature of the sample before he starts to work
12. During data collection, post notes announcing that data acquisition for a potentially hazardous sample is on going
13. Wipe the NMR tube with Kimwipes before loading the sample to minimize transfer of dirt to the probe
14. If you generate garbage (e.g. used paper towels, Kimwipes, parafilms, or any clutter in general), throw them in regular waste bin
15. Clean your working area before leaving the lab
    

Injury logs and reporting

1. In case of fire or magnet quench, call the University police, Dr. Sukumaran and Dr. Szyperski.
2. If you cut or puncture yourself while handling potentially biohazardous biofluids, wash the ezposed area with soap and water and seek medical attention from Milliard Filmore Suburban Hospital or from Health Services as soon as possible.
3. Report the incident to Dr. Szyperski, Dr. Ashan and Dr. Sukumaran
4. You should also contact EH&S (sharps injury reporting)
5. Fill out an accident report form (http://hr.buffalo.edu/files/phatfile/Workers_Comp.pdf) and follow the procedures therein
6. Record the incident in ‘Sharp Injury Log book’

---

NMR

Sample Storage and Preparation

Preparation of 1.1929611141M stock formate  

1. Combine 26.4mL 88% formic acid = 22.593961M (using pipet A) with 473.6mL D₂O + 4.50008g NaCl (filled to 500mL mark of volumetric flask +/- 0.15mL) 
   • 1.192961141M formate, D2O, 0.9% NaCl

Preparation of 27mM formate

1. Combine 11.63mL 1.192961141M stock formate (using pipet B) with 500mL D₂O with NaCl using volumetric flask +/- 0.2mL
   • ~510mL 27.11752256mM formate / D2O, 0.9% NaCl

Fetal Bovine Serum (FBS)

1. Thaw FBS at 4C overnight.
2. Original Procedure:
   1. Combine 10mL 27mM formate in D₂O with 990mL FBS
      • 1.0L FBS with 0.27mM formate 0.9% NaCl
3. Modified to:
   1. Combine 10mL 27mM formate with 90mL D₂O
      • 100mL 2.7 mM formate
      1. Combine 100mL 2.7 mM formate with 900mL FBS
         • 1L FBS with 0.27mM formate

Human Serum

Microflow Probe

Each serum specimen was thawed at room temperature inside a hood and prepared by combining 27 μL of serum and 3 μL of ‘lock solution’ (27 mM formate dissolved in D2O containing 0.9 % NaCl). The resulting solution was filtered through a barrier tip (catalogue # 87 001-866; VWR International, West Chester, PA, USA) into a 12 x 32 mm glass screw neck vial (Waters Corp., Milford, USA) by centrifugation. Additionally, 10 μL of the thawed serum was snap-frozen in liquid Nitrogen then stored at -80^oC until needed for MS analysis. The remaining thawed serum was snap-frozen and stored at -80^oC until needed for cryogenic probes data collection.

Cryoprobe

The serum was thawed at room temperature in a hood and prepared by combining 119 μL serum and 51 μL ‘lock solution’ (D2O containing 0.9 % NaCl) and transferred to a NMR thick-walled tube (inner diameter = 2.2 mm). The specimens were snapfrozen in the NMR tube after data collection and stored at -80^oC.

Data Acquisition

Profiling of serum

Microflow Probe

I. The proper functioning of the instrumentation is verified in five different steps:

1. Flow cell and path are cleaned and checked for proper functioning,
2. B0-field homogeneity is optimized by manual shimming,
3. the sample temperature (T = 25.0 ± 0.2 oC) is confirmed by recording a 1H spectrum for methanol,
4. the 1HDO signal line width (LW) and line shape detected for 99.96% D2O are assessed (acceptable values: at 50% of signal maximum< 1 Hz; at 10% < 3 Hz; at 0.55% < 10 Hz; at 0.11% < 20 Hz),
5. the S/N value of the anomeric 1H signal detected for a 10mM sucrose standard solution is measured (acceptable S/N ≥ 40).

II. The system performance for data collection for bio-fluid is evaluated in three steps:

1. A 1D 1H-NMR spectrum is recorded with 32 scans for a sample containing 27 mM of the internal standard formate dissolved in D2O containing 0.9% NaCl, after (i) the probe is tuned, and (ii) the 90o pulse width is determined, so that the LW and S/N of the formate 1H signal can be measured (acceptable values: S/N ≥ 150:1; LW at 50% of signal maximum < 1.2 Hz; at 10% < 5 Hz; at 0.55% < 15 Hz; at 0.11% < 30 Hz),
2. A sample of sterile-filtered fetal bovine serum (FBS) containing ‘lock solution’ (9:1 v/v, the ‘lock solution’ contains 27 mM of the internal standard formate dissolved in D2O containing 0.9% NaCl) is loaded in to the probe. The NMR probe is tuned, a 1D 1H-NMR spectrum without pre-saturation is recorded and the LW of the 1HDO signal is assessed (acceptable values: at 50% of signal maximum < 30 Hz; at 10% < 70 Hz), 3. for the FBS sample, the 90o 1H pulse width is determined and saturation parameters (saturation ‘carrier’ 1H frequency and saturation power) are optimized by recording standard 1D 1H NMR spectra with pre-saturation of the residual water line. The minimal saturation power is chosen which ensures that the intensity of the residual water line is lower than the intensity of the intensity of the internal standard.

III. The NMR data collection for biological specimens is pursued in eight steps:

1. The serum/plasma specimens are prepared as described in section 1.1.6.2.1,
2. The specimens are loaded in the auto sampler,
3. The NMR acquisition parameters obtained with the sterile FBS sample (see II. above) are used to set-up 1D ‘NOESY’ and 1D ‘CPMG’ 1H NMR data acquisitions,
4. The characteristics of the formate resonance line are assessed for the spectrum recorded for the first specimen (acceptable values: S/N ≥ 120:1 after 256 scans ; line width at 50% of signal maximum <1.5 Hz; at 10% < 5 Hz; at 0.55% < 15 Hz; at 0.11% < 30 Hz),
5. If the values obtained in step III.4 are acceptable, data collection continues for all remaining specimens, with D2O being used as the ‘push solvent’ and the probe being cleaned between recording of spectra with 2% Zymit™-H2O (Catalogue number Z-9701-12; International Products Corporation, Burlington, NJ), and rinsed with D2O,
6. After completion of data collection, proper spectrometer performance is confirmed by repeating step II.1 (see above),
7. The probe is cleaned using a solution of 1 M KOH in ethanol, formic acid/acetonitrile and 2% Zymit™-H2O followed by a D2O rinse,
8. The quality of all spectra is assessed by repeating step III.4 for each of the spectra

Cryogenic Probe

I. The proper functioning of the instrumentation is verified in four different steps:

1. The VT settings ( VT unit = 15^oC, VT air = 15, VT cooling air = 20 ) are verified and the temperature is measured accurately by recording a 1D 1H NMR spectrum for a 10 mM TmDOTP5- sample after which the chemical shift (relative to TSP) of the H6 is determined (acceptable value: -151.93±0.03 ppm) ,
2. B0-field homogeneity is optimized by manual shimming with a 1% CHCl3-acetone-d6 sample. Monitor the improvement of lineshape by overlaying the spectra (acceptable values: at 50% of signal maximum < 0.6 Hz; at 0.55% < 8.1 Hz; at 0.11% < 18.0 Hz). Note: Make sure that the probe is ‘detuned’,
3. The S/N value of the quartet around 1600 Hz detected for 0.1% ethyl benzene-CDCl3 and measured at 1200 – 4000 Hz spectral range is assessed (acceptable value: 1300:1),
4. The (S/N) value of the anomeric 1H signal comprising two scalar coupling fine structure components detected for a 2 mM sucrose standard solution is measured (acceptable S/N ≥ 140:1) within the 3000Hz – 5000 Hz spectral range. Additionally, the linewidth for each signal component and minimum between the two signal components are assessed (typical values: 1.5 Hz and < 20% from the baseline, respectively).

II. The performance for data collection for bio-fluid is evaluated in seven steps:

1. A 1D 1H-NMR spectrum is recorded with 4 scans for the ‘lock solution’ (contains 2.7 mM of the internal standard formate dissolved in D2O containing 0.9% NaCl) after (i) the probe is tuned, (ii) shimmed, and (iii) the 90^o pulse width is determined, so that the line width and S/N of the formate 1H signal can be measured (acceptable values: S/N ≥ 200:1; line width at 50% of signal maximum < 0.8 Hz; at 0.55% < 10.8 Hz; at 0.11% < 24.0 Hz),
2. The 90^o 13C pulse width is determined for a 13C labeled Histidine sample (50mM histidine in D2O containing 0.9% NaCl) using the first increment of the HSQC experiment. In a ‘gChsqc’ experiment, array ‘calH’ (for 1H) prior to arraying the calC (for 13C). During the 90^o 13C pulse calibation, the δ-CH signal is used to determine the ‘null’. The expected 90^o 13C pulse width at pwClvl of 59 dB is 14.2 μs, 3. (i) the probe is tuned and (ii) shimmed for a sample of sterile-filtered fetal bovine serum (FBS) to which ‘lock solution’ (7:3 v/v) was added,
3. B0-field homogeneity is optimized by manual shimming by recording a standard 1D 1H NMR spectra with pre-saturation of the residual water line until the line width for each of the lactate 1H doublet is <1.3Hz,
4. For the FBS sample, the 90^o 1H pulse width is determined and saturation parameters (saturation ‘carrier’ 1H frequency and saturation power) are optimized by recording standard 1D 1H NMR spectra with pre-saturation of the water line,
5. Acquire a 1D 1H NMR spectrum with pre-saturation of the water line with 64 scans and check that formate signal meets the specifications in II.1. The expected S/N for the formate is 120:1with no applied window function. Note: FBS contains added formate with a 0.27 mM concentration plus the biological formate. Thus, the S/N is higher than expected,
6. Using the FBS sample, 90^o 1H pulse width is determined and saturation parameters (saturation ‘carrier’ 1H frequency and saturation power) are optimized for the 1D DOSY and DIRE using their respective pulse sequence. These parameters are optimized for 2D [1H,1H] TOCSY and 2D [13C,1H] HSQC using the first increments of each experiment.

III. The consistency of set-up is assessed in two steps:

1. The NMR acquisition parameters obtained with the FBS sample (see II. above) are used to set-up 1D ‘NOESY’ and 1D ‘CPMG’ 1H NMR data acquisitions for an FBS sample,
   
2. After completion of data collection, the spectra are processed and principal component analysis (PCA) is performed with this data and the data obtained from the certification of spectrometer operators.

IV. The NMR data collection for biological specimens is pursued in six steps:

1. The serum specimen is prepared,
2. The NMR probe is tuned and shimmed after the specimen is loaded. Steps II.3 to II.5 is repeated after calibration of the 90o 13C pulse width,
3. The NMR acquisition parameters obtained with the FBS sample (see II. above) are used to set-up 1D ‘NOESY’, 1D ‘CPMG’, 1D ‘DOSY’, ‘DIRE’, 2D J-resolved 1H NMR data acquisitions as well as the 2D [1H,1H] TOCSY and 2D [13C,1H] HSQC. Additional 1D ‘NOESY’ data acquisition is set-up after 2D [13C,1H] HSQC to check if technical and specimen variability occurs during the data acquisition,
4. The line width characteristics of the formate resonance line are assessed for the 1D ‘NOESY’ and 1D ‘CPMG’ 1H NMR spectra recorded (acceptable values with LB = 0.25: line width at 50% of signal maximum <1.2 Hz; at 0.55% < 16.2 Hz; at 0.11% < 36 Hz); and the quality of water suppression is assessed by measuring the height and ‘footprint’ of the residual water line (typical values: peak maximum < intensity of the formate resonance line; < 180 Hz, respectively),
5. After 5 specimens are run (with the previous spectra included), PCA is performed. All subsequent spectra were confirmed if they populate within the 95% ellipse of the PCA plot. This step identifies any outlier,
6. After data collection is finished for 5-7 specimens, an FBS sample is inserted at random to check the set-up consistency and data integrity by repeating step III.

Data Processing

1D 1H NMR spectra

NOESY and CPMG

Processing of the 1D NMR spectra involves the iterative use of Varian VNMRJ (for the zero-filling, manual phase correction and first-order baseline correction) and Chenomx NMRSuite software (for spectral chemical shift calibration and binning). The linebroadening (LB) value to be utilized for a given project must be selected based on a compromise between S/N for the internal standard and resolution. The optimal value for the LB is selected by plotting S/N vs LB and overlay of the spectra for several values of LB.

In VNMRJ:

1. In expX, load the FID, apply the desired line-broadening value, zero-fill to 131072 points then Fourier transform
2. Adjust the phase interactively (perform zero-order and first-order phase correction at least 3X), note down the zero (called ‘rp’ in VNMRJ) and first order (called ‘lp’) phases
3. Define the noise regions to be used for the baseline correction (14.78-9.00, 5.65-5.45 and -1.00- -5.22 ppm)
4. Perform a linear baseline correction by executing the command ‘bc(1)’
5. Visually check if the baseline averages out to zero by placing the ‘threshold’ line at the middle of the corrected baseline, repeat processing if needed
6. Exit from expX (this is necessary for the next steps)
7. On the command window, go to the spectrum folder (***.fid) and copy the ‘datdir’ folder (i.e. processing parameters) from expX by performing the command:

cp -r /home/vnmr1/vnmrsys/expX/datdir

In Chenomx:

1. Copy and open the VNMRJ processed spectrum in the ‘Processor’ module
2. Visually recheck the baseline if it averages out to zero by ‘blowing up’ the noise regions, reprocess the spectrum in VNMRJ, if needed
3. Calibrate the chemical shift to the formate line at 8.444 ppm
4. Perform a phase correction interactively if the residual line between the actual and model formate line is not zero, note down the zero and first order phase corrections (phase corrections in this step, if any, should not exceed ± 30)
5. Calculate Figures-Of-Merit (FOMs) for phase correction, formate peak symmetry and baseline correction. Values should NOT exceed those in the table below:
   
6. Export new phase values into VNMRJ and repeat A
7. Confirm figures of merit for phase, baseline correction and formate peak symmetry

	Phase	Symmetry	Baseline
upper limits	residual/totalformateArea (%)	diff/totalformateArea (%)
average	±5	±8	X 10-6
stdev	±3	±4	X 10-6

DOSY and DIRE

The same data processing procedure as for 1D 1H NOESY and CPMG was used for 1D 1H DOSY and DIRE except for spectral calibration. For these diffusion-edited spectra, the most isolated resonance line for choline +N(CH3) moiety of phospholipids at 3.20 ppm was used for calibrating each spectrum.

2D NMR spectra

Homonuclear 2D J-resolved

No zero-filling was performed in both dimensions. Sinebell and exponential window functions were applied to the direct dimension while sine-bell was applied to the indirect dimension prior to Fourier transformation (FT). A first-order baseline correction was performed then a 45o tilt was done to make the J-coupled multiplets (F1 axis) orthogonal to the F2 axis (Figure 2.2.2). The spectrum was corrected for baseline again and symmetrized about the J=0 Hz. The 2D spectra was calibrated to formate resonance line at (8.444, 0.000) ppm. These processing steps were executed using NMRPipe.

The skyline projection of the 2D J-resolved spectrum was generated by repeating the same processing steps in VNMRJ prior to calculating the projection along ω1 (Figures 2.4.2 and 2.4.3).

2D [1H,1H]-TOCSY

The time domain data were zero-filled to 16384 and 512 points along F2 and F1 dimensions, respectively. Cosine-bell squared window functions were applied to both dimensions prior to FT. Each 2D spectrum was calibrated to the formate resonance line at (8.444, 8.444) ppm after phase and baseline correction.

2D [13C,1H]-HSQC

The time domain data were zero-filled to 16384 and 2048 points along F2 and F1 dimensions, respectively. A cosine-bell squared window function was applied to F2 and a cosine-bell window function was applied to F1 dimension prior to FT. Each 2D spectrum was calibrated to the formate resonance line at (8.444, 172.0) ppm after phase and baseline correction.

Outlier Identification

The initial criteria for selection of outliers is based on S/N and line width for the internal standard formate, quality of water suppression and spectral processing artifacts. This is followed by a more thorough outlier identification steps by Principal Component Analysis (PCA) scores plot and Distance-to-Model plot. The details are shown below.

Before Multi Variate Data Analysis (MVDA)

1. After each batch of data acquired, perform a quality assessment of each 1D NMR spectrum (as indicated in SOP for serum sample data acquisition). Exclude the spectra which have unacceptable formate H1 resonance S/N and linewidth. Additionally, exclude spectra with unacceptable water suppression.
2. Overlay all 1D-NOESY (or 1D CPMG) spectra and expand the formate region to check if the calibration is accurate. Scroll over the overlay to check for obvious phase and baseline errors, and by rephrasing, recalibrating, and/or repeating the baseline correction.
3. Identify spectra with anomalous (e.g. very intense and/or broad) peaks and spectra with peaks on unexpected positions. Check spectra individually and judge if each is to be excluded.

After MVDA

1. Plot PC1 vs PC2 (Pareto-scaled, bin size = 0.005ppm) and generate a Distance-to-Model (DModX) plot. Tabulate the observations (samples) which are outside the 95%-ellipse. Rank them according to distance from the ellipse. Repeat this for DModX.
2. List the observations which are classified as outliers based on both 95%-ellipse and DModX.
3. Repeat steps (1)-(2) on PCs 4 to 10 (This number of PCs considered is for a large number of samples N) pairwise.
4. Check the score contribution plot of each observation on the table generated according to (2). The contribution plot (Figure 3.2.8.A) displays the differences, in scaled units, for all the variables in the model, between a specific spectrum and the average spectrum, multiplied by the absolute value of the normalized weight (SIMCA-P+ Manual.Version 11.0. 2005. UMETRICS, Sweden). In other words, the contribution plot of an observation explains why it is different from most of the samples. List the bins with the most significant contribution (most intense).
5. For each sample, check the peak(s) corresponding to the bin(s) identified in (4) on the processed 1D NMR spectrum. If the peak(s) is(are) isolated and judged as anomalous (refer to Figure 3.2.8.B), replace the bin intensities in the excel/csv file with the median value (of other spectra). Otherwise, exclude the spectrum (Figure 3.2.8.C shows an example of an ‘outlier’ excluded for MVDA).
6. Repeat steps (4) – (6) for the rest of the observations on the table from (2). Using the modified excel/csv file, redo the MVDA. Repeat outlier identification process after performing MVDA.
7. Record and report all potential outliers and justification for exclusion, modification or neither. The summary should indicate which spectra are deemed as outliers due to the presence of resonance lines of unknown identity and for reasons stated in (1 above).

Identification of metabolites and associated pathways that are affected upon onset of a disease

After NMR data collection, processing and statistical analysis, the information from NMR and statistics are interpreted into the context of biological changes in the human body upon onset of a disease. The step-wise procedure towards building a metabolic hypothesis from changes in the NMR profile of metabolites present in serum is outlined below.

1. Determine the spectral regions of 1D 1H DIRE, 1D 1H DOSY, skyline projection of 2D Jresolved, 1D 1H CPMG, 1D 1H NOESY, 2D J-resolved and 2D [1H,1H]-TOCSY spectra that are significantly different (p-value < 0.05) between spectra recorded for specimens from ‘healthy controls’ and the ‘diseased’ group.
2. Perform resonance assignment for the spectral regions identified in step (1). Metabolites with p-values < 0.05 will be referred to as ‘significant’ metabolites from here onwards.
3. Determine which metabolites in step (2) increase or decrease in spectra of ‘diseased’ relative to ‘healthy controls’.
4. Color-code and rank the ‘significant’ metabolites according to p-values. Use the following colors to indicate degree of significance: black ≤ 10^–5, violet = 10^–4, green = 10^–3 and orange = 0.05 – 0.01.
5. Map the ‘significant’ metabolites onto human metabolic pathways which can be obtained online from the Kyoto Encyclopedia of Genes and Genomes (KEGG; http://www.genome.jp), starting with those that are colored black and violet prior to those that are colored green and orange.
6. Complement the resulting metabolic chart from step (5) with information on the location of pathways perturbed with the onset of the disease and the observations regarding tumor metabolism from literature. The site of each metabolic pathway (e.g., organelle, cell, tissue, organ) can be obtained from the Human Metabolome Database (HMDB; http://www.hmdb.ca/).
7. Compare the changes in NMR profile of metabolites in serum with the metabolite profile for the disease obtained from tissue/tumor biopsies. This could be used to determine if the changes in NMR profiles observed in serum is a systemic response of the human body to the damaged organ/tumor.

Spectral Quality Assessment (FOM)

Calculation of Figures-Of-Merit (FOM) for processed 1D NMR spectra

To ensure (i) unbiased baseline correction, (ii) a high quality of shims and (iii) an accurate phase correction, FOMs were calculated for each spectrum. For (i), the average and standard deviation of data points on the spectral regions containing only baseline (e.g., 14.7–9.0 ppm and –1.0 – -5.2 ppm) were calculated. The cut-offs chosen for the EOC project were: average < 2 x 10–6 and standard deviation < 5 x 10–6. For (ii) and (iii), a Lorentzian line was fitted into the resonance line of the internal standard formate. The residual (that is, the difference between the actual peak shape and the Lorentzian line) was integrated after the fit. The cut-offs chosen were <2% of the total integral of the formate resonance line. Additionally, the symmetry of the formate line was assessed by calculating the difference between peak segments located down- and upfield of the peak maximum. The cut-off chosen for the difference was <4% of the total integral.

FOM Calculations for Phase Correction

1. In Chenomx profiler:  fit the formate line at 8.444ppm and save with residual line.
2. Bin the data for subtraction line and spectrum line with standardized area.
   • 8.0 - 9.0ppm with bin size of 0.00031ppm
   • Exclude region: 8.00-8.4368 and 9.45208-9.00ppm
3. FOM for phase correction:
   • Sum of absolute value of residual formate peak / Sum of the total points of formate peak * 100%

FOM Calculations for Phase Correction

1. FOM of baseline correction:
   • Average of points defining the baseline / Sum of total points of formate peak * 100%
2. CPMG from -5 to 14.5ppm
   • Bin at 0.005ppm

FOM Calculations for Peak Symmetry

1. FOM for peak symmetry
   • Difference between two sides of formate peak / sum of total points of formate peak * 100%
2. Bin the data for spectrum line with standardized area.
   • 8.00 to 9.00ppm with bin size of 0.00031ppm
   • Exclude region: 8.00-8.4368 and 9.45208-9.00ppm

Operator Certification

It is important to assess not only the variability arising from the specimen collection, sample preparation, NMR data collection and processing but also to evaluate the variability arising from the operator pursuing the SOP. The operator-associated variability can be used to decide if one is ‘certified’ to pursue NMR-based metabonomics research, or if further training is required.

1. Read the Magn. Reson. Chem. 2009 paper (Sukumaran, DK; Garcia, E; Hua, J; Tabaczynski, W; Odunsi, O; Andrews, A; Szyperski, T. Standard operating procedure for metabonomics studies of blood serum and plasma samples using a 1H-NMR micro-flow probe. Magn. Reson. Chem. 2009, 47, S81-85) as an introduction to the certification procedure.
2. Perform the necessary checks to ensure proper functioning of the instrumentation (see SOP for 1H NMR data acquisition, Section I) and evaluate performance of the spectrometer for biofluids (see SOP for 1H NMR data acquisition, Section II.
3. Acquire 1D 1H NOESY and 1D 1H CPMG spectra for fetal bovine serum (FBS) (or any other test samples provided there is ample number of spectra (at least about 10 spectra) acquired by a certified operator on the same spectrometer and the test samples you want to use were prepared as one batch with those used by the certified operator) that were prepared in the same batch as those that were used by the certified operator. Also, make sure that you use the same type of NMR tube as the ones used by the certified operator. The trainee should collect spectra for at least 5 FBS samples per set up.
4. Perform spectral quality assessment (see SOP for 1H NMR data acquisition, Section IV.4) after data collection for each sample.
5. Process the 1D 1H NMR spectra (see SOP for data processing). Perform a PCA using SIMCA-P for your processed spectra and the spectra of a certified operator.
6. Repeat steps 2 – 5 at least 3X wherein someone else uses the spectrometer in between your set ups.
7. Certification is granted if all spectra from all set ups are indistinguishable from those acquired by the certified operator (i.e., your spectra should be scattered around the spectra of the certified operator within the 95 % ellipse of PCA score plots, see Figure 3.2.2 for example).
8. Bin (bin size = 0.005 ppm, unless other specification is asked for) your processed spectra together with the spectra acquired by the certified operator and send it to the statisticians (Prof. Chris Andrews and Mr. Su Qian, Department of Biostatistics, University at Buffalo) for additional analysis (if necessary) and for certification confirmation.

Resonance Assignment

Chemical shifts information from literature, databases and Statistical Total Correlation Spectroscopy (STOCSY)59 were utilized for the resonances observed in 1D 1H NOESY and 1D 1H CPMG. Confirmation of resonance assignment was obtained by use of homonuclear 2D J-res, 2D [1H,1H]-TOCSY and 2D [13C,1H]-HSQC at natural abundance and by spiking. The SOP for resonance assignment is described in the following:

1. After processing the spectra (i.e. zero-filling to 131,072 points, line-broadening for exponential window function= 1.2 Hz, FT, phase correction baseline correction, spectral calibration to α-Glucose H1 resonance at 5.233 ppm), select a 1D CPMG spectrum for resonance assignment. This spectrum should contain the resonance peaks corresponding to the most influential bins/variables in the PCA loadings. Resonances of the 1D CPMG spectrum will be assigned based on literature41,60-62 starting from the most intense and well-resolved peaks followed by the less intense, well-resolved peaks. Resonance lines in the crowded regions will be tentatively assigned based on the published chemical shifts and multiplicities.
2. Generate a 1D STOCSY59 plot for the 1D CPMG spectra to determine peaks which are positively correlated. Verify from literature41,60-62 if the correlated peaks belong to the same molecule. Repeat this process until all correlated peaks are considered. Peaks which cannot be assigned at this point will be labeled as unassigned resonances and they will be grouped according to their correlation coefficient obtained from 1D STOCSY.
3. On a 1D NOESY spectrum, carry over the assignment from steps 1 and 2. For the remaining peaks (e.g. broad peaks) start the assignment based on literature41,60-62 starting from the intense, well-resolved lines going to the crowded regions.
4. Repeat step 2 using the 1D NOESY data.
5. Group peaks as ‘assigned’ or ‘unassigned’ peaks. The unassigned resonances shall be grouped further according to their correlation coefficients in STOCSY.
6. Confirm assignments by 2D J-resolved, 2D [1H,1H] TOCSY and 2D [13C,1H] HSQC (natural abundance). Additionally, confirm/obtain correlations for the unassigned peaks.
7. For the unassigned peaks, prioritize the assignment by:
   1. Loading contribution in PCA
   2. S/N
8. Perform a database search for unassigned resonances (http://mmcd.nmrfam.wisc.edu/, http://www.metabolibrary.ca/, http://www.bmrb.wisc.edu/metabolomics/query_metab.php, NMRSuite 5.0 (Chenomx, Canada, etc.).
9. Rank the candidate compounds by the number of information obtained from STOCSY, 2D J-resolved , 2D [1H,1H]-TOCSY, 2D [13C,1H]-HSQC (natural abundance) and peak integrals.
10. Spike the compounds as ranked in 9 to obtain ultimate confirmation of resonance assignments.

Data Archiving

Since metabonomics involves handling of a large number of raw and processed data, it is essential to maintain a good ‘bookkeeping’ and data back-ups.

1. Take the sample list (i.e. excel file) from the Sample Storage Protocol and record the raw FID name for each NMR experiment along with the run order numbers.
2. Record the filename of the corresponding processed spectra beside the raw FID column.
3. When generating processed spectra for multivariate data analysis (in the form of N x K matrix), indicate the project name and whether the spectra are binned or unbinned in the filename. Record the data description in your lab journal. Additionally, generate a list of sample filename and info/description for each specimen and upload it to the HTP server (/nsm/chem/cen2/HTP2/1_projects/METABONOMICS).
   1. After data collection, the raw FID should be saved in the spectrometer computer and lab desktop as soon as possible.
   2. The raw FID and the processed spectra should be saved in at least one of the lab desktops (e.g. meta (NSC 820), moscow (NSC 820) and spins3 (NSC 807)), HTP server and a separate disk.