Gut microbiota and SCFAs have been shown to have a wide range of beneficial physiological effects on the host throughout the lifespan [39, 40]. Even though several studies have characterized the microbial profile in adults, there is scarce literature describing the evolution of the bacterial ecosystem during early life beyond weaning period. This work provides a picture of how the bacterial composition and excretion of SCFAs changes upon the introduction of complementary feeding within the first year of life, supporting the importance of breastfeeding in the modulation of the infant's microbiota composition and activity.
From birth, the gut microbiota develops rapidly into a succession of bacterial strains  in a transition to an adult microbiota. Throughout this process different factors, related to pregnancy and delivery and the infant's environment, influence the establishment of the different bacterial groups. Among them type of delivery, gestational age, gender, or type of early feeding are known to have an influenced and, then, they should be considered when interpreting diet-microbiota relationships [5, 41].
The difference observed in the Bifidobacterium mean level between breastfed and formula-fed infants at 15 days of age (9.17 ± 1.46 vs. 8.76 ± 1.33, data not shown) is consistent with the ability of this bacteria to degrade some HMOs from human milk such as lacto-N-biose [42, 43]. The first colonizers of the newborn's gastrointestinal tract are aerotolerant microorganisms which reduce oxygen levels facilitating an optimal environment for the proliferation of anaerobic bacteria such as Bifidobacterium [42, 44]. According with that, bifidobacteria have been shown as the most abundant genus in the sample up to 6 months of age  and declined during the weaning process . In line with previous studies in Spanish children the first food groups introduced in the sample during the weaning period were infant cereals (64.1% of the sample), fruits (55.1%), vegetables (44.9%) and tubers (43.6%) .
The reason for the changes on the microbiota following complementary feeding are not totally understood yet. Some authors argue that cessation of breastfeeding may play a more important role than the introduction of solid foods [6, 48, 49], but based on our results it seems reasonable to speculate that the incorporation of complex polysaccharides in the infant´s diet, may favor the growing of some bacteria groups, such as Bacteroides and Clostridium cluster IV, which are specialized in the breakdown of fiber [50–52]. In this line, although given the nature of the study we are unable to establish the directionality, our results pointed to a direct association between Clostridium cluster IV levels and dietary fiber throughout the first year of life. Some authors have reported that diets enriched in different type of fiber, like inulin, oligofructose, guar gum, arabinoxylan and resistant starch, induce Clostridium cluster IV enrichment . Also, the decrease in Enterobacteriaceae and Bifidobacterium observed during the first months of life contrary to what occurs with Bacteroides and Clostridium cluster IV at 12 are in agreement with previous longitudinal studies .
Acetate is the main SCFAs in the sample, as occurs in adults . The increase of Bacteroides and Clostridium groups with the weaning process may probably determine the increment in acetate and butyrate observed from 6 to 12 month [19, 51] since some strains of Clostridium genus are able to transform acetate into butyrate [55, 56].
Finally, the classification of infants, at 6 and 12 months, according to their fecal microbiota and short-chain fatty acid resulted in two cluster: being cluster I represented by higher excretory infants in comparison to cluster II. Perhaps one of the mayor findings of this study is the identification of at breastfeeding as one of the major determinants of cluster membership. The higher levels of propionate and butyrate acids showed in formula-fed are in agreement with some of the first studies investigating this relationship . With independence of breastfeeding, it is possible that, undigested proteins and amino acids can pass into the large intestine, being fermented by certain bacteria producing different bacterial metabolites, such as SCFAs [58, 59]. Consistent with this finding, we observed that infants belonging to cluster I (those with higher excretion of SCFAs) had significantly higher total protein intake than cluster II at 90 and 180 days (10.36 g/day vs. 8.75 g/day and 15.17 vs. 13.47, respectively) (data not shown).
Some authors have hypothesized that SCFAs represent the key molecular link between diet, microbiome and health playing an important role throughout the lifespan in protecting the body against deteriorating metabolic control and inflammatory status . In this regard, while some studies related the high excretion of SCFAs with obesity risk , we have not observed significant differences in infant body weight among clusters. However, those children belonging to cluster I (higher SCFAs excretion and infant formula intake) experienced more weight gain from birth to 3 and 6 months than those belonging to cluster II (2.71 vs. 2.24 kg and 4.70 vs. 4.20 kg, respectively) (data not shown). Further research is needed to determine the normal range of SCFAs levels and the threshold above which they may have an impact on health.
The study has some limitations related to its observational nature and the collection of dietary data. In interpreting this information, it should be noted that the energy and nutrient content of processed infant foods has been taken into account, a factor that has hitherto been underestimated in the literature. Regarding the quantification of breast milk energy, it is necessary to mention a limitation of the study. Since it was not possible to record the exact volume of milk produced by the mother, an indirect estimation was made using the mean amounts established in the literature for each age range [30, 31]. While the quality of the FFQ depends on the respondent's memory, its ability to accurately classify energy and all nutrient intakes in children is enhanced by the fact that the questionnaires are adapted from the PANCAKE study and have photographs that make them easier to interpret. In addition, it allows comparison with other studies on the European infant population.