1. Introduction

To assess the relevance of the Firmicutes/Bacteroidetes ratio as a taxonomic signature of healthy, we used data from microbiota composition from nine published studies carried out in seven countries (USA, United Kingdom, India, Pakistan, Chile, Argentina, and Colombia) and including 728 healthy subjects. High-throughput 16S rRNA gene sequence data from previous studies, corresponding to V3–V4 or V4 hypervariable regions and generated by the Illumina MiSeq Platform, were collected (Table 1). To allow direct comparisons among sequences from different studies, all the reads were filtered using the DADA2 pipeline, then aligned and trimmed to the same length (80 bp) using Mothur, followed by the taxonomic identification using the DADA2 pipeline based on the identification of Exact Sequence Variants ^[1].

2. Heterogeneity

We considered that a sequence length of 80 bp was adequate for analyzing the microbial communities at the Phylum level (less than 0.02% of the reads were not assigned). Using a unique pipeline and reads generated by a same sequencing platform, we expected to eliminate all bias generated by sequencing and bioinformatic tools, as commented above. Subsequently, we analyzed the relative abundance of Firmicutes and Bacteroidetes and their ratio. Overall, the data indicate that the abundance of Firmicutes in the gut microbiota of healthy individuals varies between 11% to 95% and that of Bacteroidetes between 0.6% to 86.6% (Figure 1). Considering this variability, it seems difficult to observe significant changes for these two phyla in obese people. In their meta-analysis, Finucane et al. observed that the variations for both Firmicutes and Bacteroidetes abundance were much larger among studies than between lean and obese individuals within any study ^[2]. In agreement with these findings, we recently observed a high variability of both Firmicutes and Bacteroidetes, between 25%–67% and 4%–64% respectively, in the fecal microbiota of young healthy Chilean volunteers, despite rigorous inclusion criteria including the control of anthropometrical and biochemical markers, biomarkers of systemic and colonic inflammation (plasma IL-6 and high sensitivity C-reactive protein and fecal calprotectin, respectively), and dietary intake ^[3]. Again, the heterogeneity of the diet is probably the main factor explaining such variations in the healthy population and this might eventually make difficult the identification of specific microbial signatures. For example, on the one hand, Wu et al. showed in 100 healthy individuals with known dietary habits that the microbiota from those consuming protein and fat-based diets were enriched with Bacteroides whereas that from those consuming carbohydrate-based diets were enriched with Prevotella ^[4], results from which these authors formulated the concept of “enterotype”. On the other hand, Balamurugan et al. compared the gut microbiota from a tribal population (Malayalis) living in the northern part of Tamil Nadu (India) that consumed a restricted diet due to cultural and religious beliefs, with that from healthy villagers from the same region as controls. Both populations exhibited a high abundance of Firmicutes (85.9% and 63.5%, for the Malaiyalis and controls, respectively) and low abundance of Bacteroidetes (2.65% and 0.45%, respectively), resulting in very high Firmicutes/Bacteroidetes ratio (34.0 and 92.9, respectively), though the individuals from both populations were lean ^[5]. Although the Malaiyalis population had a restricted, homogenous diet, a high variation in the proportions of both Firmicutes and Bacteroidetes was observed, confirming that factors other than diet influence this ratio.

Table 1. Description of studies considered in this study for evaluating the variability of the Firmicutes and Bacteroidetes.

^& Effective obtained after the bioinformatic processing; * Data obtained from publishing data (not-recalculated due to the lack of individual data); ** Data submitted for publication, provided by Susan Pesoa, co-author of this work; UK ¹, UK ², Argentina ¹ and Argentina ² are studies reported in the Figure 1.

Figure 1. Variability in the Firmicutes/Bacteroidetes ratio (A) and the relative abundances of Firmicutes (B) and Bacteroidetes (C) in the gut microbiota from several healthy populations. Box plots were constructed using R. In the box and whisker plots, the line shows the median; the box, the interquartile range; and the whiskers, the highest and lowest values.