ASICS: an automatic method for identification and quantification of metabolites in complex 1D 1H NMR spectra

Patrick J.C. Tardivel1,4, Cécile Canlet1,2, Gaëlle Lefort3, Marie Tremblay-Franco1,2, Laurent Debrauwer1,2, Didier Concordet1, Rémi Servien1

1Toxalim, Université de Toulouse, INRA, ENVT, INP-Purpan, UPS, F-31027 Toulouse, France.

2Axiom Platform, MetaToul-MetaboHUB, National Infrastructure for Metabolomics and Fluxomics, F-31027 Toulouse, France.

3GenPhySE, Université de Toulouse, INRA, ENVT, CastanetTolosan, France

4 Corresponding author: , +33561193995

Results obtained on a reference human plasma: NIST Standard Reference Material for human plasma (SRM1950).

Plasma samples were received frozen and stored in their original vials at -80°C until use. Prior to use, an ampule was thawed on ice and vortexed for 30 s. 500 µl of plasma sample were mixed with cold methanol in a 1:2 ratio (v/v), vortexed, and incubated at -20°C for 20 min. The mixture was centrifuged at 11000 rpm for 30 min to pellet proteins. Supernatants were decanted to fresh vials and dried under nitrogen. The dried sample was mixed with 600 µL of phosphate buffer prepared in deuterated water (0.2 M; pH 7.0) and the mixture was transferred into 5 mm wide NMR tube.

Theplasma 1H NMR spectrum was recorded using the cmpgpr1d NMR sequence for the suppression of macromolecules and water resonances, with a spectral width of 20 ppm, 3 s relaxation delay, with 128 transients collected into 32k data points.

The 10 - 0.5 ppm range of the spectrum was used, excluding the 5.1 - 4.5 ppm region for water signal.

Three competing methods were tested on this plasma. Results are presented in the following table together with the identification performed by Simón-Manso et al (2013) and the quantitative data generated manually by an NMR expert. As mentioned in the main text, Chenomx identified 78 metabolites whereas Bayesil identified 44 metabolites. In a sack of compactness, all these metabolites are not reported here. We only report the metabolites that have been identified by ASICS and/or the NMR expert.

To confirm the chemical structure of metabolites, 2D 1H-1H COSY (Correlation Spectroscopy) and 2D 1H-13CHSQC (Heteronuclear Single Quantum Coherence spectroscopy) NMR experiments were performed on the NIST plasma sample.

Four different levels of metabolite identification have been defined by Summer et al. (2007), identified metabolites (level 1), putatively annotated compounds (level 2), putatively characterized compound classes (level 3), and unknown compound (level4). For level 1, a minimum of two independent and orthogonal data relative to an authentic compound analyzed under identical experimental conditions are proposed as necessary to validate non-novel metabolite identification (full 1H NMR spectra and 2D NMR spectra). All the identifed compounds at level 1 are pointed on Figure S1. We only report the metabolites identified by ASICS or/and by the NMR expert.

Compound / PubChem CID / Metabolite proportion
(ASICS) / Metabolite proportion
(NMR expert) / Metaboliteproportion (Bayesil) / Metabolite proportion (Chenomx) / Identification by Simon-Manso et al / Identification level (NMR expert)
D-Glucose / 5793 / 1 / 1 / 1 / 1 / yes / 1
L-Lactic acid / 107689 / 0.6404 / 0.233 / 0.330 / 1.099 / yes / 1
L-Glutamine / 5961 / 0.1195 / 0.089 / 1.245 / 0.181 / yes / 1
L-Alanine / 5950 / 0.0884 / 0.067 / 0.064 / 0.222 / yes / 1
Glycine / 750 / 0.0611 / 0.040 / 4.019 / 0.113 / yes / 1
L-Proline / 145742 / 0.0570 / 0.094 / 0.117 / - / yes / 1
L-Valine / 6287 / 0.0551 / 0.040 / 0.157 / 0.104 / yes / 1
L-Threonine / 6288 / 0.0364 / 0.0171 / 0.485 / 0.071 / yes / 1
L-Leucine / 6106 / 0.0358 / 0.024 / 0.155 / - / yes / 1
L-Lysine / 5962 / 0.0336 / 0.053 / - / - / yes / 1
D-Fructose / 2723872 / 0.0327 / - / - / - / no / N.I.*
3-hydroxybutyric acid / 441 / 0.0322 / 0.030 / 0.051 / 0.072 / yes / 1
Glycerol / 753 / 0.0321 / 0.066 / 2.171 / - / yes / 1
Acetic acid / 176 / 0.0305 / 0.017 / 0.009 / 0.098 / yes / 1
L-Serine / 5951 / 0.0274 / 0.067 / 3.394 / - / no / 1
Taurine / 1123 / 0.0258 / - / - / - / no / N.I.*
L-Histidine / 6274 / 0.0249 / 0.011 / 0.626 / - / yes / 1
L-Glutamic acid / 33032 / 0.0247 / 0.019 / 0.0458 / - / yes / 1
D-Glucuronic acid / 94715 / 0.0233 / - / - / - / yes / N.I.*
Trimethylamine-N-Oxide / 1145 / 0.0225 / - / - / 0.0031 / no / N.I.*
Glycerophosphocholine / 439285 / 0.0223 / 0.013 / - / 0.050 / no / 1
L-Isoleucine / 6306 / 0.0217 / 0.012 / - / 0.043 / yes / 1
Maltose / 439186 / 0.0216 / - / - / - / no / N.I.*
Ascorbic acid / 54670067 / 0.0213 / - / - / - / no / N.I.*
D-Mannose / 18950 / 0.0196 / 0.0064 / - / 0.090 / yes / 1
S-Acetamidomethylcysteine / 1590100 / 0.0191 / - / - / - / no / N.I.*
L-Tyrosine / 6057 / 0.0179 / 0.010 / - / - / yes / 1
Trans-4-hydroxy-L-Proline / 5810 / 0.0178 / - / - / - / no / N.I.*
L-Aspartic acid / 5960 / 0.0173 / - / - / - / no / N.I.*
L-Phenylalanine / 6140 / 0.0167 / 0.0091 / 0.118 / - / yes / 1
Ethanolamine / 700 / 0.0143 / - / - / - / no / N.I.*
Hypotaurine / 107812 / 0.0126 / - / - / - / no / N.I.*
Creatinine / 588 / 0.0119 / 0.0097 / 0.068 / 0.0215 / yes / 1
L-Ornithine / 6262 / 0.0117 / - / - / 0.0325 / no / N.I.*
Phosphocholine / 1014 / 0.0117 / - / - / - / no / N.I.*
1,6-Anhydro-Beta-D-Glucose / 0.0112 / - / - / - / no / N.I.*
Choline / 6209 / 0.0092 / 0.0184 / 0.0475 / - / yes / 1
D-Fucose / 840 / 0.0089 / - / - / - / yes / N.I.*
L-Glutathione-oxidized / 65359 / 0.0084 / - / - / - / no / N.I.*
3-methyladipic acid / 12292 / 0.0068 / - / - / - / no / N.I.*
Pantothenic acid / 6613 / 0.0066 / - / - / 0.0015 / no / N.I.*
Glycogen / 439177 / 0.0059 / - / - / yes / N.I.*
N,N-dimethylglycine / 673 / 0.0016 / 0.00043 / - / 0.0020 / no / 12
L-Asparagine / 6267 / - / 0.0098 / 0.0344 / 0.0311 / no / 12
Creatine / 586 / - / 0.0052 / - / 0.020 / no / 1
Betaine / 247 / - / 0.021 / 0.016 / 0.0044 / no / 1
L-Tryptophan / 6305 / - / 0.0083 / 0.124 / 0.023 / no / 1
Formic acid / 284 / - / 0.0186 / - / 0.105 / no / 1

Table S1. Comparison of ASICS, Bayesil, Chenomxand a NMR expert quantification and the identification of Simón-Manso et al (2013) on the NIST plasma.

Figure S1. 600 MHz 1H NMR spectrum of deproteinized NIST plasma. (A) 0.5 to 4.5 ppm; (B) 5.1 to 9 ppm. The region B has been magnified two times as compared to the region A. Numbers correspond to the following metabolites : 1, Isoleucine; 2, Leucine; 3, Valine; 4, 3-hydroxybutyric acid; 5, Lactic acid; 6, Threonine; 7, Alanine; 8, Lysine; 9, Acetic acid; 10, Proline; 11, Glutamine; 12, Glutamic acid; 13, Creatine; 14, Creatinine; 15, Choline; 16, Glycerophosphocholine; 17, Glucose; 18, Betaine; 19, Glycine; 20, Glycerol; 21, Serine; 22, Mannose; 23, Tyrosine; 24, Histidine; 25, Phenylalanine; 26, Tryptophan; 27, Formic acid.

* N.I. = not identified

References

Simón-Manso, Y., Lowenthal, M.S., Kilpatrick, L.E., Sampson, M.L., Telu, K.H., Rudnick, P.A., Mallard, W.G., Bearden, D.W., Schock, T.B., Tchekhovskoi, D.V., Blonder, N., Yan, X., Liang, Y., Zheng, Y., Wallace, W.E., Neta, P., Phinney, K.W., Remaley, A.T., & Stein, S.E. (2013). Metabolite Profiling of a NIST Standard Reference Material for Human Plasma (SRM 1950): GC-MS, LC-MS, NMR, and Clinical Laboratory Analyses, Libraries, and Web-Based Resources. Analytical Chemistry,85(24), 11725-11731.

Sumner, L.W., Amberg, A., Barrett, D., Beale, M.H., Beger, R., Daykin, C.A., Fan, T.W.-M., Fiehn, O., Goodacre, R., Griffin, J.L., Hankemeier, T., Hardy, N., Harnly, J., Higashi, R., Kopka, J., Lane, A.N., Lindon, J.C., Marriott, P., Nicholls, A.W., Reily, M.D., Thaden, J.J., & Viant, M.R. (2007). Proposed minimum reporting standards for chemical analysis Chemical Analysis Working Group (CAWG) Metabolomics Standards Initiative (MSI). Metabolomics, 3, 211–221.