banner

Blog

Jun 19, 2024

Inter

ISME Communications volume 3, Numéro d'article : 84 (2023) Citer cet article

945 accès

15 Altmétrique

Détails des métriques

La recherche sur les communautés microbiennes marines se développe, mais les études sont difficiles à comparer en raison des différences dans les protocoles d’échantillonnage de l’eau de mer. Pour aider les chercheurs à comparer les études utilisant différentes méthodologies d’échantillonnage de l’eau de mer, ainsi que pour les aider à concevoir de futures campagnes d’échantillonnage, nous avons développé l’initiative EuroMarine Open Science Exploration (EMOSE). Dans le cadre d'EMOSE, nous avons échantillonné des milliers de litres d'eau de mer provenant d'une seule station du nord-ouest de la Méditerranée (Service d'Observation du Laboratoire Arago [SOLA], Banyuls-sur-Mer), pendant une seule journée. L'ensemble de données résultant comprend plusieurs approches de traitement de l'eau de mer, englobant différents types de filtres (membrane à cartouche et membrane plate), trois fractionnements de tailles différentes (> 0,22 µm, 0,22 à 3 µm, 3 à 20 µm et > 20 µm), et un certain nombre de volumes d'eau de mer différents allant de 1 L à 1000 L. Nous montrons que le volume d'eau de mer filtré n'a pas d'effet significatif sur la diversité des procaryotes et des protistes, indépendamment de la stratégie de séquençage. Cependant, il y avait une nette différence de diversité alpha et bêta entre les fractions granulométriques et entre celles-ci et « l’eau entière » (sans pré-fractionnement). Dans l’ensemble, nous recommandons d’être prudent lors de la fusion de données provenant d’ensembles de données utilisant des filtres de tailles de pores différentes, mais nous considérons que le type de filtre et le volume ne doivent pas agir comme des variables confondantes pour les stratégies de séquençage testées. À notre connaissance, c’est la première fois qu’un ensemble de données accessible au public permet effectivement de clarifier l’impact des options méthodologiques du microbiome marin sur un large éventail de protocoles, y compris des variations à grande échelle du volume échantillonné.

La caractérisation de la vie microbienne sur Terre est devenue un sujet d’intérêt transdisciplinaire. En effet, il est désormais reconnu qu’il est crucial d’acquérir des connaissances en écologie microbienne dans plusieurs biomes pour développer une compréhension plus approfondie de la vie, des cellules aux écosystèmes. En conséquence, de vastes projets de recherche collaboratifs internationaux se sont concentrés sur les microbiomes associés aux humains [1, 2], aux coraux [3], aux herbiers marins (https://seagrassmicrobiome.org/protocols/) ou aux éponges [4, 5]. En outre, au cours des 20 dernières années, des initiatives coordonnées à l’échelle mondiale d’échantillonnage du microbiome planctonique marin ont été lancées, telles que le Global Ocean Sampling (2003-2010) [6, 7], le Recensement international des microbes marins (ICoMM) [8], Expédition de circumnavigation Malaspina 2010 [9] et expéditions Tara Ocean (2009-2012) [10], ainsi que des programmes de recensement tels que le programme Earth Microbiome [11, 12] et la Journée d'échantillonnage des océans (OSD) dirigée par Micro B3 [13 ]. Les détails sur les avancées et les perspectives sur l’écologie microbienne des océans mondiaux, leur pertinence et les défis futurs ont été largement examinés ailleurs [14].

L'effort massif actuel pour étudier les microbiomes du monde a donné lieu à de multiples initiatives de normalisation, notamment l'utilisation de protocoles communs pour l'échantillonnage du microbiome de différents environnements et tissus hôtes, ainsi que de procédures de séquençage communes. Les initiatives pertinentes avec des efforts de standardisation méthodologique incluent, par exemple, OSD [13], Earth Microbiome [15], European Marine Omics Biodiversity Observatory Network (EMO BON) [16] et Metagenomics for Human Intestinal Tract (MetaHIT) [17].

L’analyse à grande échelle des microbiomes libres et associés à l’hôte constitue un énorme pas en avant dans la compréhension des interactions microbes-animaux [3, 5, 18, 19] et microbes-plantes [20], ainsi que de la structure, de la fonction et diversité des communautés microbiennes dans divers habitats terrestres [21]. Cependant, des lacunes doivent être comblées pour mieux standardiser et harmoniser les meilleures pratiques et stratégies d’échantillonnage, de description et d’étude de la diversité microbienne. En particulier, dans les études sur le microbiome océanique, il est connu que l’estimation de la richesse microbienne dépend de plusieurs facteurs, notamment les gènes marqueurs et les amorces utilisés pour le métabarcoding [22], les différents protocoles d’extraction d’ADN [23, 24], la profondeur de séquençage et les paramètres génomiques. approche (séquençage des amplicons vs séquençage du métagénome) et critères de clustering (25). Bien qu’il soit reconnu que la stratégie d’échantillonnage influence les estimations de la diversité microbienne du plancton [25], il existe un manque d’études conçues pour tester systématiquement l’effet des variables méthodologiques sur les procédures d’échantillonnage pour étudier la diversité du microbiome océanique et la composition taxonomique. Ces études sont cruciales pour concevoir des protocoles précis pour échantillonner toute la gamme de tailles des communautés microbiennes marines (26).

0.22 µm) and of the 0.22 µm to 3 µm size fractions used 142 mm diameter polyethersulfone Express Plus membrane filters (Product Code GPWP14250, Millipore). For the 3 µm to 20 µm fractionations, 142 mm diameter polycarbonate membrane filters were used (Product Code TSTP14250, Millipore). As for the large size fractions (>20 µm), the 47 mm diameter nylon mesh filter was used instead (referred to as flat membrane from here on)./p>20 µm). Additionally, whole water cartridge membrane volumes from 1 L to 10 L were also compared for the metagenomes. Below, we consider the prokaryotes and protists results independently./p>0.22 µm) or size fractions (0.22–3 µm, 3–20 µm and >20 µm) in columns. Color distinguishes between flat and cartridge membrane filters. Within each grid unit, the prokaryotic species richness is plotted against volume, which ranges from 2.5 L to 1000 L./p>0.22 µm) and 0.22–3 µm size fraction samples presented a similar number of prokaryotic taxonomic lineages (Fig. 3a) and both presented fewer prokaryotic taxonomic lineages than the 3–20 µm size fraction (Fig. 3a). Accordingly, the statistical test indicated significant differences in the species richness obtained after > 0.22 µm, 0.22–3 µm and 3–20 µm (p < 0.05, Kruskal–Wallis), more specifically, between >0.22 µm and 3–20 µm size fractions (p < 0.05, post-hoc Dunn test). On the metagenomes side, for the same comparison, there were no appreciable differences in the number of prokaryotic taxonomic lineages (Fig. 3a) and they were not significant (p > 0.05, Kruskal–Wallis). Details on the above-mentioned statistical tests are available in Supplementary Table S7./p>0.22 µm), 0.22–3 µm and 3–20 µm size fractions for the same volume (10 L) and filter (flat membrane), for MetaB16SV4V5 (left) and metagenomes (right). Note that metagenomes didn’t include samples in 3–20 µm size fraction in (a). b Comparison for size fractions (0.22–3 µm, 3–20 µm and > 20 µm size fractions) for the same volume (100 L) and filter (flat membrane), for MetaB16SV4V5 (left) and metagenomes (right). Note that metagenomes didn’t include samples in >20 µm size fraction in (b). c Comparison for flat membrane vs cartridge membrane, for the same volume (10 L) and whole water (>0.22 µm), for MetaB16SV4V5 (left) and metagenomes (right). d Comparison between 2.5 L (single filter) and 10 L (four 2.5 L filters pooled together), using the same filter (cartridge membrane) and whole water (> 0.22 µm), for MetaB16SV4V5 (left) and metagenomes (right). All panels illustrate the species richness obtained for each sample (point). To help the reader compare the variables, we added boxplots on top of the points. Significance was determined using either Mann–Whitney test for two independent groups, or Kruskall–Wallis for more than two independent groups, followed by a post-hoc Dunn test, if needed. Significance was illustrated with the symbols: p > 0.05 (empty); p < 0.05 (*); p < 0.01 (**); and p < 0.001 (***)./p>20 µm, using the flat membrane filter, which revealed an increase in the prokaryotic species richness with increasing pore size, for both MetaB16SV4V5 and metagenomes (Fig. 3b). In fact, the median number of prokaryotic taxonomic lineages obtained by MetaB16SV4V5 increased significantly from 335 (0.22–3 µm) to 429 (3–20 µm) and 538 (>20 µm) (p < 0.05, Kruskal–Wallis, Fig. 3b), more specifically between 0.22–3 µm and > 20 µm size fractions (p < 0.05, post-hoc Dunn test). Similarly, metagenomes increased the median number of prokaryotic taxonomic lineages from 155 (0.22–3 µm) to 195 (3–20 µm) (Fig. 3b), which was also significant (p < 0.05, Mann–Whitney). Details on the above mentioned statistical tests are available at Supplementary Table S7. Please note that for metagenomes in Fig. 3b there are no samples for the >20 µm size fraction because some samples were lost due to insufficient DNA for sequencing, while some samples that were successfully sequenced were later discarded due to low number of reads (below 10 000 reads, for a list of discarded samples in the rarefaction step see Supplementary Table S4). The overview of prokaryotic species richness was overall consistent and supported by the rarefaction curves because the different size fractions had similar levels of alpha diversity, while the same did not apply for volume (Supplementary Figs. S2 and S3)./p> 0.05, Mann–Whitney). Metagenomes provided an equivalent number of prokaryotic taxonomic lineages between either filter (Fig. 3c) and the differences were not significant (p > 0.05, Mann–Whitney). Although we compared cartridge and flat membrane filters under the same volume (10 L), the cartridge membrane filters reached 10 L by pooling together four cartridge membrane filters of 2.5 L. However, the single 2.5 L cartridge membrane filter and 10 L pooled from four cartridge membrane filters of 2.5 L obtained an equivalent number of prokaryotic taxonomic lineages, without significant differences (p > 0.05, Mann–Whitney) for either sequencing approach (Fig. 3d). Details on the above mentioned statistical tests are available at Supplementary Table S7./p>20 µm size fractions. Additionally, the volume did not follow any clear direction in the ordination figures (Fig. 4a, b). PERMANOVA tests were made to support the ordination figures, with similar results for MetaB16SV4V5 and metagenomes. Specifically, both volume and size fractions significantly changed the community composition (p < 0.05, PERMANOVA), but this result should be interpreted with caution, because if the same test considers the division of samples by size fraction, then community composition did not change significantly across volume (p > 0.05, PERMANOVA). Details on the PERMANOVA statistical tests for prokaryotes are available at Supplementary Table S8. The variation within size fractions, measured by distance to centroid, further supported the clustering of prokaryotic community composition by size fractions (Fig. 4c,d, Supplementary Table S9)./p>0.22 µm), 0.22–3 µm, 3–20 µm and >20 µm size fractions. Division by (a) MetaB16SV4V5 and (b) metagenomes. Additionally, boxplots represent the distance to centroids of samples within each size fraction, divided by (c) MetaB16SV4V5 and (d) metagenomes. Note that metagenomes didn’t include the >20 µm size fraction. For details on missing replicates, we refer the reader to Supplementary Table S1./p> 20 µm) using the same volume (100 L) and filter (flat membrane)./p>0.22 µm) for MetaB18SV9 showed any appreciable change in the protist species richness from 2.5 L (median = 343, IQR = 6.75, n = 4) to 10 L (median = 348, IQR = 34.8, n = 12) (Fig. 7). However, for either MetaB18SV9 and metagenomes, there was no appreciable difference in the protist species richness from 10 L to 1000 L, within any of the size fractions (Fig. 7). Comparing pore sizes, whole water (>0.22 µm), 3–20 µm and >20 µm size fractions identified more protist taxonomic lineages than 0.22–3 µm size fraction samples (Fig. 7). The number of protist taxonomic lineages obtained after each sample are available at Supplementary Table S10. The higher impact of size fraction, rather than volume, on protist species richness was further supported by rarefaction curves (Supplementary Figs. S2 and S3), even though the size fractions were not as distinct from one another as they were with the prokaryotic data./p>0.22 µm) or size fractions (0.22–3 µm, 3–20 µm and >20 µm) in columns. Color distinguishes between flat membrane and cartridge membrane filters. Within each grid unit, the protist species richness is plotted against volume. For details on missing replicates, we refer the reader to Supplementary Table S1./p>0.22 µm) or 3–20 µm size fraction (Fig. 8a). However, the range of the number of protist taxonomic lineages obtained for whole water included the range of values for both the 0.22–3 µm and 3–20 µm size fractions (Fig. 8a). More specifically, the number of protist taxonomic lineages obtained by MetaB18SV9 varied between 290 and 380 for the whole water, 289 and 338 for 0.22–3 µm size fraction, and 338 to 357 in 3–20 µm size fractions (Fig. 8a), which were not significantly different (p > 0.05, Kruskal–Wallis). The number of protist taxonomic lineages obtained by metagenomes varied between 88 and 129 for whole water, 91 and 97 for 0.22–3 µm, and 105 and 128 for 3–20 µm size fractions (Fig. 8a); these differences were also statistically non-significant (p > 0.05, Kruskal–Wallis). We note, however, the number of samples for the metagenome provide little support for the described differences in this specific comparison. Details on the above mentioned statistical tests are available in Supplementary Table S11./p>0.22 µm), 0.22–3 µm and 3–20 µm size fractions for the same volume (10 L) and filter (membrane), for MetaB18SV9 (left) and metagenomes (right). b Comparison for size fractions (0.22–3 µm, 3–20 µm and > 20 µm size fractions) for the same volume (100 L) and filter (membrane), for MetaB18SV9 (left) and metagenomes (right). c Comparison for flat membrane vs cartridge membrane, for the same volume (10 L) and whole water (>0.22 µm), for MetaB18SV9 (left) and metagenomes (right). d Comparison between 2.5 L (single filter) and 10 L (four 2.5 L filters pooled together), using the same filter (cartridge membrane) and whole water (> 0.22 µm), for MetaB18SV9 (left) and metagenomes (right). All panels illustrate the species richness obtained for each sample (point). To help the reader compare the variables, we added boxplots on top of the points. Significance was determined using either Mann–Whitney test for two independent groups, or Kruskall–Wallis for more than two independent groups, followed by a post-hoc Dunn test, if needed. Significance was illustrated with the symbols: p > 0.05 (empty); p < 0.05 (*); p < 0.01 (**); and p < 0.001 (***)./p>20 µm size fractions for the same filter (membrane) and volume (100 L), the 0.22–3 µm size fraction had fewer protist taxonomic lineages than the 3–20 µm and >20 µm size fractions (Fig. 8b), for either MetaB18SV9 and metagenomes. These differences were significant for the MetaB18SV9 (p < 0.05, Kruskal–Wallis), but not for the metagenomes (p > 0.05, Kruskal–Wallis). However, the significance of the test was not very strong and the post-hoc test for MetaB18SV9 was not significant for any combination of size fractions, after adjustment (p > 0.05, post-hoc Dunn test). Details on the above mentioned statistical tests are available in Supplementary Table S11./p>0.22 µm). The differences in the number of protist taxonomic lineages between cartridge and flat membrane filters were small (Fig. 8c) and not significant (p > 0.05, Mann–Whitney). However, the range of values was wider for the flat membrane filter than the cartridge membrane filter with the MetaB18SV9 approach (Fig. 8c). The number of protist taxonomic lineages within the replicates of flat membrane filters varied between 290 and 380 (difference of 90 taxonomic lineages), while in the cartridge membrane filters varied between 354 and 373 (difference of 19 taxonomic lineages) (Fig. 8c). For metagenomes, the values were equivalent between both types of filters (Fig. 8c). Please note that the cartridge membrane and flat membrane filters were compared at 10 L volume, but the cartridge membrane samples obtained 10 L by pooling together four cartridge membrane filters of 2.5 L together. For MetaB18SV9, the number of protist taxonomic lineages obtained after pooling four 2.5 L cartridge membrane filters was higher than using a single filter of 2.5 L (Fig. 8d), but not significant (p > 0.05, Mann–Whitney). However, this was not the same for the metagenomes, where the number of protist taxonomic lineages was equivalent and slightly higher for a single filter of 2.5 L (Fig. 8d), but also not significant (p > 0.05, Mann–Whitney). Details on the above mentioned statistical tests are available at Supplementary Table S11./p>20 µm size fractions were distant from the remaining, in either MetaB18SV9 and metagenomes (Fig. 9a, b). This was further supported by the significant results of PERMANOVA for the volume and size fractions independently (p < 0.05, PERMANOVA), but once they were considered together the effect on community composition was no longer significant (p > 0.05, PERMANOVA). Note that the variable for size fractions did not meet the homogeneity of variance pre-requisite of PERMANOVA (p > 0.05, betadisper). Details on the PERMANOVA statistical tests for protists are available in Supplementary Table S12. Additionally, a more detailed look into the betadisper results, i.e., a measure of distance to the centroid of samples within each size fraction, revealed that samples were very consistent within size fractions (Fig. 9c,d and Supplementary Table S13)./p>0.22 µm), 0.22–3 µm, 3–20 µm and >20 µm size fractions divided by (a) MetaB18SV9 and (b) metagenomes. Additionally, boxplots represent the distance to centroids of samples within each size fraction, divided by (c) MetaB16SV4V5 and (d) metagenomes./p>20 µm size fraction consistently identified more taxonomic lineages, independently of the volume, for example, Dinophyceae, Bacillariophyceae and Foraminifera (Fig. 10). In contrast, other groups were more prevalent in the 3–20 µm size fraction, like Cercozoa, Hacrobia and Haptophyta (Fig. 10). Several groups did not seem to favor any specific size fraction, like Excavata or Syndinales (Fig. 10). In the metagenomes, from 10 L to 1000 L, some groups had more protist taxonomic lineages in the >20 µm size fraction, like Bacillariophyceae, or fewer, like Hacrobia (Supplementary Fig. S5). Additionally, the metagenomes did not reveal any specific taxonomic group that increased the number of protist taxonomic lineages with increasing volume (Supplementary Fig. S5)./p>20 µm size fraction consistently had more taxonomic lineages, indicating that several taxonomic lineages were specifically found in that size fraction. One possible explanation for the identification of taxonomic lineages specific to the >20 µm size fraction is that those prokaryotes were attached to particles, or to the filter material itself. Considering that the turbidity of the water was very low, the only particles plausible for the prokaryotes to attach to would be the protists or other cell debris, including aggregates. Thus, we suggest that the prokaryotic taxonomic lineages specific to the large size fraction could be prokaryotes associated with microeukaryotes, colonial bacteria and/or specialized in colonizing larger particles. Given the presence of prokaryotes on > 20 µm size fractions and protists on 0.22–3 µm size fractions, we cannot rule out the possibility that extracellular DNA, besides actual cells, is retained in the filters, for example, by sorption [64]. However, the general picture is that free-living prokaryotes are identically identified in whole water (> 0.22 µm) and 0.22–3 µm size fraction, while particle-attached prokaryotes can be retained within larger pore size fractions (3–20 µm and >20 µm). This is consistent with previous studies that account for the effect of pre-filtration on prokaryotic diversity with 16 S rRNA gene sequencing [65]. Protists also follow the same general picture described in previous studies [40], with contamination between smaller size fractions, for example, because of cell fragments. In this study, either biological group was most unique in composition at >20 µm size fraction. Notwithstanding, we highlight that it was unexpected to find more prokaryotic and protist taxonomic lineages in the > 20 µm size fraction than in whole water, which cannot be fully explained by our experimental design and should be addressed in future work./p>0.22 µm) was generally equivalent to the 0.22–3 µm size fraction. This metabarcoding and metagenomic comparison of sampling protocols can help researchers to design their own sampling campaigns and to compare studies using different protocols. Even though we did a tremendous effort to address many different variables in protocols used by different campaigns, there is more to be tested and compared for the purpose of standardization of protocols in the future, for example, DNA extraction protocols./p>

PARTAGER