*2.6. Study Variables*

Body weight, BMI, and free fat mass were measured using bioelectrical impedance analysis (BIA) on a whole body BIA analyzer (Tanita BC-420MA, Tanita Corporation, Tokyo, Japan). Biochemical analyses included urinary isoprostanes (8-iso-PGF2α, ELISA kit, Oxford Biomedical Research, Rochester Hills, MI, USA), serum malondialdehyde (MDA) (MDA ELISA kit, Elabscience, Houston, TX, USA), and serum oxidized low-density lipoprotein (Ox-LDL) (Human OxLDL ELISA kit, Elabscience) as lipid-related oxidative stress biomarker; urinary 8-hydroxy-2- -deoxiguanosine (8-OHdG) (ELISA kit, Elabscience) as DNA-related oxidative stress biomarker, and serum protein carbonyl (Protein Carbonyl ELISA kit, Enzo Life Sciences, Lausanne, Switzerland) as protein-related oxidative stress biomarker; and serum glutathione peroxidase (GPx) (ELISA kit, Elabscience) and serum superoxide dismutase (SOD) (ELISA kit, Elabscience) as endogenous antioxidative enzymes. Safety analyses included complete blood count, liver function tests (bilirubin, alanine and aspartate aminotransferases, gamma-glutamyl transpeptidase), and renal function tests (blood urea nitrogen and serum creatinine levels).

For microbiome analysis, DNA was isolated with the aid of a QIAmp Power Fecal Pro DNA kit (Qiagen, Hilden, Germany), with bead beating and enzymatic lysis steps prior to extraction to avoid bias in DNA purification toward misrepresentation of Grampositive bacteria. Massive genome sequencing of the hypervariable region V3–V4 of the bacterial 16s rRNA gene was conducted to evaluate the bacterial composition of the gut microbiome. Samples were amplified using key-tagged eubacterial primers [22] and sequenced with a MiSeq Illumina Platform, following the Illumina recommendations for library preparation and sequencing for metagenomic studies. The resulting sequences were split per patient, considering the barcode introduced during the PCR reaction. R1 and R2 reads were overlapped using PEAR program version 0.9.1, with an overlap of 50 nucleotides and a quality of overlap with a minimum of Q20, providing a single FASTQ file for each of the samples. Quality control of the sequences was performed by initial quality filtering (minimum threshold of Q20) using fastx tool kit version 0.013, followed by primer (16s rRNA primers) trimming and length selection (reads over 300 nts) with cutadapt version 1.4.126. These FASTQ files were then converted to FASTA files, and chimeras that could arise during the amplification and sequencing steps were removed by the UCHIME program, version 7.0.1001. Those clean FASTA files were BLAST against the National Center for Biotechnology Information (NCBI) 16s rRNA database using blastn version 2.2.29+. The resulting XML files were processed using a python script developed by ADM-Biopolis; (Valencia, Spain) to annotate each sequence at different phylogenetic levels.
