Samplestat ver 3

The peaks detected by CE-TOFMS and LC-TOFMS were processed with Master Hands ver.2.13.0.8.h (Keio University) to obtain m/z values, peak areas, and migration time (CE-TOFMS) or retention time (LC-TOFMS)^{21 (link)}. The metabolic pathway map was provided using the public-domain software VANTED (Visualization and Analysis of Networks containing Experimental Data, Germany)^{22 (link),23 (link)}. Principal component analysis (PCA) and hierarchical clustering analysis (HCA) were performed using SampleStat ver.3.14 and PeakStat ver.3.18 (Human Metabolome Technologies). PCA is the most widely used dimension-reducing technique for analysing the large datasets involved in metabolome analysis^{24 (link)}. A heatmap was generated for the HCA, with red and green indicating high and low concentrations, respectively. All values were presented as the mean ± standard deviation. We considered a p value less than 0.05 to be significant (Welch’s t test).

Body weights during weeks 3–10 in Experiments 1 and 2 were analyzed by repeated two‐way analysis of variance (ANOVA) with group and age as factors. In Experiment 1, body weight, epididymal fat, and testicular weight, as well as composition of muscle fiber types and muscle genes expression at 10 weeks of age were analyzed by two‐way ANOVA with photoperiod during 0–4 and 4–8 weeks of age as factors, followed by a Bonferroni multiple comparison test. In Experiment 2, body weight, epididymal fat, and testicular weight at 10 weeks of age were analyzed by one‐way ANOVA followed by a Bonferroni multiple comparison test.
Metabolome data were analyzed using principal component analysis (PCA, SampleStat ver. 3.14, Human Metabolome Technologies) and hierarchical cluster analysis (HCA, PeakStat ver. 3.18, Human Metabolome Technologies). Heat maps were generated by coloring the values of all data according to the color scale. Relative area values were used for calculation of the ratio of 0–4 LD to SDC groups and for comparison between 0–4 LD and SDC groups using Student's t‐test. Values were considered significantly different at P < 0.05.

Frozen mice liver samples (n = 2 in the HF group, n = 3 in the other groups) were transferred into 500 µL of methanol containing 50 µM of external standard. After homogenization by BMS-M10N21 (bms, Tokyo) at 1,500 rpm, 120 s five times, 500 µL of chloroform and 200 µL of ultra-pure water were added to the homogenate and mixed well and centrifuged at 2,300 g for 5 min at 4°C. The resultant water phases were ultrafiltrated by the Millipore Ultrafree-MC PLHCC HMT Centrifugal Filter Device, 5 kDa (Millipore, Billerica, MA). The filtrates were dried and dissolved in 50 µL of ultra-pure water.
We then subjected the samples obtained to capillary electrophoresis time-of-flight mass spectrometry (CE-TOFMS) analysis using the Agilent CE-TOFMS system (Agilent Technologies, Santa Clara, CA) at 4°C. The alignment of detected peaks was performed according to the m/z value and normalized migration time. We performed a principal component analysis (PCA) with the detected peaks using the statistical analysis software SampleStat ver.3.14 (Human Metabolome Technologies Inc., Tsuruoka, Japan), and we performed a hierarchical clustering analysis (HCA) using PeakStat ver.3.18 (Human Metabolome Technologies). The relative area value of each peak was calculated and used for the intergroup comparison. Samples that were obviously characterizing outliers were eliminated from the analysis.