Machine learning of plasma metabolome identifies biomarker panels for metabolic syndrome: findings from the China Suboptimal Health Cohort

Table 2 Differential metabolites identified between MetS participants and controls

Metabolites	MetS group		Control group		Trend	FC	VIP	P value	P* value	P^# value
Metabolites	Mean	SD	Mean	SD	Trend	FC	VIP	P value	P* value	P^# value
Trans-Acotinic Acid	0.0043	0.0031	0.0054	0.0022	Down	0.7917	4.4239	3.543E–03	2.515E–06	2.138E–05
Methanol	0.0051	0.0012	0.0049	0.0008	Up	1.0432	1.0652	3.836E–04	9.162E–03	2.685E–02
Guanidinoacetate	0.0039	0.0011	0.0036	0.0007	Up	1.0579	1.0384	5.379E–04	9.941E–04	4.970E–03
Hydroxyproline	0.0027	0.0020	0.0035	0.0016	Down	0.7738	3.4554	2.930E–06	1.113E–06	1.052E–05
Pyroglutamic Acid	0.0021	0.0009	0.0025	0.0009	Down	0.8438	2.0394	1.155E–11	2.459E–08	5.225E–07
Glutaconic Acid	0.0008	0.0010	0.0014	0.0010	Down	0.5850	2.6766	3.108E–15	4.362E–10	1.236E–08
D-Maltose	0.0014	0.0011	0.0019	0.0008	Down	0.7296	2.2688	1.547E–09	8.209E–08	1.395E–06
D-Fucose	0.0010	0.0005	0.0008	0.0003	Up	1.2343	1.2169	3.648E–10	1.209E–11	1.028E–09
Taurine	0.0017	0.0006	0.0014	0.0005	Up	1.1609	1.5231	1.675E–08	4.260E–04	2.481E–03
Deoxyadenosine	0.0002	0.0004	0.0004	0.0004	Down	0.6409	1.0993	2.990E-06	4.379E–04	2.481E–03
L-Ornithine	0.0010	0.0013	0.0013	0.0013	Down	0.7367	1.8037	3.846E–04	1.291E–03	5.777E–03
L-Asparagine	0.0007	0.0006	0.0009	0.0006	Down	0.7888	1.0710	2.398E–06	4.334E–03	1.635E–02
Carnosine	0.0003	0.0004	0.0004	0.0005	Down	0.6662	1.0302	1.743E–05	3.734E–04	2.442E–03

Down trend, relatively lower levels of metabolites present in MetS group. Up trend, relatively higher levels of metabolites present in the MetS group
MetS metabolic syndrome, FC fold change, VIP variable importance on projection
P value, P value from Wilcoxon test without adjustment
P < 0.05 was considered statistically significant
P^* value, P value adjusted for age
P^# Value, P value adjusted for age and false discovery rate using the Benjamini‒Hochberg method

ISSN: 1475-2840