Gut microbiota profiles of young South Indian children

Gut microbiota has been implicated as a modifier of childhood growth. Here, 16S rRNA sequencing-based fecal microbiota profiles of 18-24 month old Indian children were evaluated (n=41), in relation to their anthropometric parameters, intestinal permeability, body composition and total energy expenditure. Pathway analyses were conducted to assess microbial functions related to stunting, underweight and wasting. The fecal microbiota was enriched in Prevotella 9, Bifidobacterium and Escherichia-Shigella. Weight, weight-for-age Z-scores (WAZ) and weight-for-length Z-scores (WLZ), along with age, acted as covariates of microbiota variation specifically in boys (n=23). Bifidobacterium longum subsp longum abundance was positively associated with WAZ while Bifidobacterium bifidum and Bifidobacterium breve abundances were negatively associated with age. The lipopolysaccharide biosynthesis pathway was upregulated in stunted (n=16) and wasted (n=8) children. Findings from this study indicate that child sex may be a critical modifier of the role of gut microbiota on childhood growth.Overall design: Fecal samples were collected from non-breastfed children, aged between 18-24 months, frozen and stored. Fecal DNA content was extracted using an in-house kit that comprised of fecal sample lysis, RNase A treatment, phenol-chloroform-isoamyl alcohol-based phase separation and precipitation of DNA with 100% ethanol. Bacterial DNA was PCR amplified with primers targeting variable region V3-V4 of 16S rRNA gene using a nested PCR strategy. Cleaned V3-V4 amplicons were used for library preparation using NEBNext Ultra DNA Library preparation kit. The obtained library was pooled and sequenced with Illumina HiSeq 2500 instrument to generate 0.5M, 250bp Paired end reads/sample. To segregate the barcodes sequenced together on the machine, Illumina's bcl2fastq v2.18 was used for demultiplexing. The default parameters of bcl2fastq were retained for the demultiplexing step which notably includes allowing one mismatch in the barcode sequence. The data quality post-demultiplexing was verified by custom scripts and found to be suitable for further analysis. These analyses were performed at MedGenome Labs Pvt Ltd, Bangalore, India. The demultiplexed sequencing reads were processed using QIIME 2 (v2019.10.0) pipeline. The paired end reads were depleted of amplicon primers using the Cutadapt plugin. Quality check was performed using demux plugin followed by denoising, chimera identification and PhiX removal and dereplication using the DADA2 plugin. Taxonomies were assigned using a naive Bayes classifier (q2-feature- classifier and classify-sklearn plugins) trained on V3-V4 sequence region extracted from the latest SILVA rRNA (16S SSU) v132 reference database using locus-specific primer sequences. Operational taxonomic units (OTUs) were defined at 99% similarity.