Fig. 1: Study design and clinical improvements after the three-week IF intervention.

a Schematic diagram of this study. For the five non-fasting days per week, meal replacement powders were arranged as a staple food for dinner, and participants were instructed to eat ad libitum (labeled as “Normal intake”). For the two discontinuous fasting days, the energy intake was set on 500–600 kCal (about 25% of participants’ normal intake), provided by meal replacement powders and protein sticks. b Body mass index (BMI), weight, and atherosclerosis index (AI) showed significant improvement (*** represents adjusted p < 0.001, paired Wilcoxon test) after the intervention. The improvements were uniform among participants with different levels of BMI. Bounds of boxes represent the first quartiles (Q1) and the third quartiles (Q3) respectively, center lines represent the median values, and the whiskers are ranged from Q1–1.5 × (Q3–Q1) to Q3 + 1.5 × (Q3–Q1). c Receiver operating characteristic (ROC) curve of random forest classifiers based on the relative abundances of metagenome-assembled genomes (MAGs) before the intervention. Area under the curve (AUC) = 0.570, 0.681, and 0.582, respectively.