This study aimed to describe the microbial community present in spices sold in open-air markets in Saint-Louis using the sequencing method. The high-throughput sequencing method used in this study allowed us to detect a broad-spectrum genomic footprint of micro-organisms, including those usually not detected by culture methods. Indeed, since the revolutionary improvement in DNA sequencing technologies, direct high-throughput analysis of the genomic DNA of an entire community without prior culture has become the most common approach, overcoming the constraints of conventional microbiological approaches [28]. In addition, from a technological point of view, MinION sequencing is mainly used to address metataxonomy [29].
Our results show the presence of many pathogenic bacteria in the spice samples studied, with their abundance varying according to the type considered. Studies in countries like Brazil have reported similar results [30]. In contrast to other studies in Vietnam, Iran, Indonesia, India, and the Netherlands, analyzed spices meet established microbiological standards and were of satisfactory quality [29–30]. These were particularly spices, herbs, and dried spices produced on a small scale.
The significant bacterial communities found in the “Local Spices or Herbs” samples analyzed compared to the others could be explained by the difference in water activity. The “Local Spices or Herbs” samples, consisting of a mixture of aromatic herbs and food condiments, represent a more favourable environment for bacterial growth than the other dried spices. Nevertheless, despite their packaging giving them “stable physicochemical characteristics”, the dehydrated products, such as Curcuma, the mixture of 7-spice and thyme, have a relatively large bacterial community (Supplements 1–4). These foods generally undergo drying processes at ambient temperature at the place of production. Moreover, in the developing countries from which they originate, harvesting and production technologies do not necessarily comply with optimal sanitary conditions [29–31]. The non-compliance with hygienic measures at all stages of the process, particularly during agricultural production, harvesting, washing and sun-drying, may explain the results of this study. Packaging hygiene and storage conditions are not always respected either, which is a probable cause of food contamination (FAO, https://www.fao.org/3/w7429f/w7429f0r.htm). All these factors lead to high levels of microbial contamination, which means these products may not be suitable for human consumption [29, 32, 33].
The presence of Escherichia coli, Salmonella spp. and Bacillus spp. in most of the analyzed samples could potentially be harmful in herbs and in various food matrices such as spices, which were the subject of this study [36]. Moreover, the bacterial spores introduced by spices can withstand various preparation processes, including heat treatment [37]. The high presence of Escherichia coli in all samples may indicate faecal contamination of various origins. In the case of fat-in-house spices, including horticultural products, faecal contamination could be irrigation water from wastewater systems used without prior biological treatment [37–42]. As for the other spices studied, faecal contamination could be due to a hygiene failure by operators handling the processing (reduction to powder, mincing, etc.), storage and transport stages.
The analysis of distance matrices containing dissimilarity information can effectively capture significant and subtle compositional differences between the samples studied [44]. Other techniques, such as the Unweighted Pair Group Method with Arithmetic Mean (UPGMA), an agglomerative hierarchical clustering method widely used in practice, can also be considered. Step by step, it combines the two closest groups or elements into a higher-level group, and the distance between the new group and any other group is calculated as the arithmetic mean distance between the elements in the different groups or clusters [45–46].
The small sample size (n = 49) in this study constitutes a limitation. Furthermore, the presence of genetic material cannot provide information on the level of bacteriological contamination. Therefore, the description of the microbial community obtained by the metagenomic approach could have been accompanied by conventional microbiological analyses to highlight the fraction of viable bacteria in the analyzed samples.