A liquid chromatography-mass spectrometry (LC-MS)-based workflow LIMeX was used to extract complex lipids and polar metabolites using a biphasic solvent system of cold methanol, methyl tert-butyl ether (MTBE), and water with multiple internal standards added during the extraction and before the LC-MS analysis.

The upper (organic) phase was used for lipidomics profiling using reversed-phase liquid chromatography (RPLC), while the bottom (polar) phase was used for metabolomics analysis using RPLC and hydrophilic interaction chromatography (HILIC).

Overall, biofluids and non-fat matrices were analyzed using LIMeX-6D, while for adipose tissues, a separate analysis of highly abundant triacylglycerols was included (LIMeX-7D).

For lipidomics profiling, two RPLC (ACQUITY Premier BEH C18 column) methods with acetonitrile/water (3:2) and isopropanol/acetonitrile (9:1) as mobile phases were used with different mobile-phase modifiers based on electrospray ionization (ESI) polarity used. Specifically, the RPLC lipidomic method in ESI(+) used ammonium formate (10 mM) and formic acid (0.1%), while the RPLC lipidomic method in ESI(–) worked with ammonium acetate (10 mM) and acetic acid (0.1%). In the case of adipose tissues (isWAT, dlWAT, BAT), separate ESI(+) analyses were conducted for minor and major (triacylglycerols) lipid species.

For metabolomics profiling, HILIC (ACQUITY Premier BEH Amide column) with acetonitrile/water (95:5) and water as mobile phases, both with ammonium formate (10 mM) and formic acid (0.125%) operated in ESI(+) and ESI(–) was used followed by RPLC (ACQUITY Premier HSS T3 column) with water and methanol as mobile phases, both with formic acid (0.2 and 0.1%, respectively) in ESI(+) and ESI(–).

The dimension of all columns used was 50 mm × 2.1 mm id with 1.7–1.8 μm particle size. Each column was connected to a VanGuard FIT cartridge (5 mm × 2.1 mm id with 1.7–1.8 μm particle size).

Data sets were acquired using a Vanquish UHPLC System (Thermo Fisher Scientific) coupled to a Q Exactive Plus mass spectrometer. Simultaneous acquisition of MS1 data at 35,000 FWHM (m/z 200) and 2 data-dependent scans at 17,500 FWHM (m/z 200) was used for all platforms.

For lipidomics platforms, normalized collision energy (NCE) was 20% for ESI(+) and 10, 20, and 30% for ESI(–) was used. For metabolomics platforms, NCE 20, 30, and 40% were used for ESI(+) and ESI(–).

MS-DIAL software was used for peak alignment followed by automated annotation using in-house retention time–MS/MS libraries and compilation of MS/MS libraries (NIST20, MassBank, MoNA) for polar metabolites, and LipidBlast MS/MS library accommodated in MS-DIAL.

JTK_CYCLE, a nonparametric algorithm for detecting rhythmic components, was used for statistical analysis to identify metabolites with circadian rhythms. In addition, dryR - a tool developed to analyze rhythmicity across multiple conditions - was also applied.

The real-time quantitative polymerase chain reaction (RT-qPCR) method was used to detect mRNA levels of clock genes using a High Capacity cDNA RT Kit (ThermoFisher), specifically in the liver (Per1, Per2, Rev-Erbα, Cry1, Bmal1, Clock), SCN (Per2, Rev-Erbα, Bmal1, Dbp, E4bp4), pancreas and dlWAT (Per2, Rev-Erbα). Diluted cDNA was amplified on LightCycler480 (Roche, Basel, Switzerland) using SYBR Select qPCR Master Mix (ThermoFisher). JTK_CYCLE and dryR were used for statistical analysis.