Researchers collected 15 one-litre samples of surface seawater, with three biological replicates for each of five filter configurations: open polyethersulfone filters with pore sizes of 0.2, 1.2, 5.0 and 8.0 micrometres and enclosed 0.2-micrometre Sterivex filters. The open filters had a filtration area of about 17 square centimetres, compared with about 10 square centimetres for the enclosed filters. Pore sizes were selected in relation to expected eukaryotic cell sizes of 10–100 micrometres and bacterial sizes of 0.1–5.0 micrometres. The study tested whether larger pores would increase the proportion of eukaryotic eDNA.

Each sample was subjected to parallel shotgun sequencing and targeted 18S metabarcoding. Shotgun libraries were sequenced on an Illumina NovaSeq 6000, while the metabarcoding workflow amplified an approximately 110-base-pair region of the 18S gene. DNA was extracted with a modified Qiagen Blood & Tissue kit protocol. The shotgun analysis produced 394.4 million raw reads, of which 96.7% were merged, yielding 190.6 million paired reads. After database matching and filtering for significance, sequence similarity, query coverage and alignment length, 9.0 million reads, or 2.3% of the paired reads, remained for further analysis.

At the superkingdom level, taxonomic composition differed between small and large pores. Filters with 0.2- and 1.2-micrometre pores had mean bacterial and eukaryotic read proportions of 63% and 28%, respectively, whereas the corresponding proportions for 5.0- and 8.0-micrometre filters were 31% and 49%. Wilcoxon tests with Benjamini–Hochberg-adjusted p-values indicated more archaeal and bacterial reads with the small pores and more eukaryotic and viral reads with the large pores. An ordination based on the presence or absence of phyla also separated the samples by pore size.

Classification of the eukaryotic subset was more limited. Among eukaryotic reads, 16.4% were assigned to a kingdom, equivalent to 5.7% of the reads searched against BLAST and 0.8% of all paired shotgun reads. Small-pore filters produced more reads assigned to Viridiplantae, while large-pore filters produced more reads assigned to Metazoa, Fungi and other eukaryotic groups. The investigators did not directly measure which forms of eukaryotic eDNA were preferentially retained or the passage of bacterial cells and cellular debris through the filters.

Shotgun sequencing identified 93 phyla in total across Archaea, Bacteria, Eukaryota and Viruses, while the targeted metabarcoding analysis detected 45 eukaryotic phyla. The eukaryotic comparison showed substantial overlap: 39 phyla were detected by both methods, six were unique to metabarcoding and nine were unique to shotgun sequencing. Shotgun sequencing therefore recovered 48 eukaryotic phyla in total, compared with 45 in metabarcoding. The method-specific detections may reflect differences in amplification, reference-database coverage and unrecorded sequence diversity, so the comparison does not establish that either method consistently detects every biological phylum present.

The researchers also compared shotgun reads assigned to genera in four marine metazoan groups with genera recorded from Denmark. Among shotgun detections, 12 of 13 polychaete genera, six of seven malacostracan genera, 11 of 18 bivalve genera and 23 of 59 actinopterygian genera were classified as Danish. Corresponding metabarcoding results were 22 of 23, three of four, six of six and five of six. The authors regarded exotic-genus assignments as more likely incorrect identifications caused by incomplete reference coverage than as genuine Danish records. Reads assigned to exotic genera were consistently less abundant than those assigned to Danish genera, particularly in large-pore samples, but the study did not establish a universal read-count threshold for separating the categories.

Important limitations constrain the findings. The experiment covered one harbour and one sampling period, with only three biological replicates per filter configuration and no negative field control. The third replicate differed from the first two across filter types, including higher bacterial abundance in several phyla and lower abundance in Discosea and Echinodermata. This pattern was consistent with short-term temporal or spatial variation in coastal eDNA, but the design could not test that explanation formally. Contamination or background DNA also could not be ruled out.

Together, the results support an association between pore size and retained taxonomic composition in the seawater samples examined. They also show that shotgun sequencing can recover broad eukaryotic diversity at the phylum level despite a low overall assignment rate. The findings do not establish a universal optimum filter pore size for eDNA sampling, and the low assignment rate, limited replication, single location and absence of field controls remain central limitations of the study.

Subject of Research: Effects of filter pore size on the taxonomic composition of shotgun-sequenced seawater environmental DNA

Article Title: Filter pore size influences taxonomic composition of retained eDNA from seawater samples—evidence from shotgun sequencing

Article References: Gómez-Repollés, A., Sigsgaard, E. E., Jensen, M. R., & Thomsen, P. F. (2026). Filter pore size influences taxonomic composition of retained eDNA from seawater samples—evidence from shotgun sequencing. Metabarcoding and Metagenomics, 10, Article e164232. https://doi.org/10.3897/mbmg.10.164232

Image Credits: AI Generated

DOI: 10.3897/mbmg.10.164232

Keywords: environmental DNA, seawater, filter pore size, shotgun sequencing, 18S metabarcoding, eukaryotes, bacteria, taxonomic composition, biomonitoring, Filter, pore, size

The study enrolled 75 vegans and 75 omnivores, all free of known chronic disease, and subjected each participant to an unusually comprehensive battery of measurements. The team performed shotgun sequencing of DNA extracted from faecal samples, which reads the genetic material of the entire gut microbial community rather than just a marker gene. They also profiled metabolites in three separate bodily compartments: serum, faeces and urine, combining targeted measurements of short-chain fatty acids and bile acids with untargeted nuclear magnetic resonance spectroscopy. On top of this, they measured systemic inflammation by quantifying circulating cytokines, and collected detailed dietary records alongside standard clinical markers such as blood glucose, cholesterol fractions and blood pressure.

The dietary contrast between the two groups was exactly what the researchers expected. Vegans consumed significantly more fibre and digestible carbohydrates and significantly less protein and fat than the omnivores. But what makes the study striking is how faithfully this dietary difference was mirrored in the biology of the gut. The vegan gut bacteriome, while showing lower alpha-diversity, a measure of how many different species coexist within a single individual, displayed a clear enrichment of bacterial taxa that tracked directly with the amount of dietary fibre each person ate. Many of these fibre-associated species scored favourably on the ZOE Microbiome Ranking 2024, an index that classifies gut bacteria according to their associations with metabolic health outcomes.

One of the more surprising findings concerned protein. Within the vegan group, bacterial diversity correlated positively with the amount of dietary protein consumed, and omnivore-enriched bacterial species that are considered favourable for metabolic health according to the ZOE ranking were more abundant in vegans with a high protein intake than in those eating less protein. In other words, the study suggests that the quality of a vegan diet, not merely the absence of animal products, shapes which microbes flourish. Vegans who manage to eat plenty of plant protein appear to recruit some of the same health-associated bacterial species that are typically enriched in omnivores, hinting that protein intake is an underappreciated lever in plant-based nutrition.

The shotgun sequencing also allowed the team to look beyond bacteria to the gut virome, the community of viruses, mostly bacteriophages, that prey on gut microbes. Vegans carried a lower abundance of virulent bacteriophages and fewer viral auxiliary metabolic genes, genetic elements carried by viruses that can reprogram the metabolism of the bacteria they infect. This quieter viral landscape in vegans may reflect a different ecological balance within the gut, one shaped by the fibre-rich, plant-forward diet and the bacterial community it sustains. Because bacteriophages can transfer genes between bacterial strains and modulate microbial population dynamics, their altered profile represents a genuine, if still poorly understood, dimension of how diet reshapes gut ecology.

Perhaps the most unexpected player in the vegan gut was Blastocystis, a single-celled eukaryotic organism, often classed as a protist, whose presence was significantly more common among vegans. Long regarded with suspicion as a potential pathogen, Blastocystis has increasingly been associated in population studies with plant-rich diets, and in this cohort its presence was linked to lower systolic blood pressure. The finding adds to a growing body of evidence that Blastocystis may be a benign or even beneficial coloniser in the context of high-fibre diets, although the authors are careful to frame this as an association rather than proof of causation.

The metabolomic data told a parallel story. In omnivores, the abundance of several gut bacteriome-derived metabolites was associated with markers of adverse metabolic and renal outcomes. Two of the most notable were trimethylamine-N-oxide, or TMAO, a compound produced when gut microbes metabolise carnitine and choline from animal foods and which has been implicated in cardiovascular disease, and p-cresyl sulphate, a protein-fermentation-derived molecule that accumulates when kidney function declines. The fact that these microbe-derived compounds tracked with unfavourable clinical markers specifically in the omnivore group provides a plausible biochemical bridge between animal-product consumption and cardiometabolic risk.

To integrate this avalanche of data, the researchers used a multiblock statistical framework known as DIABLO, which combines the bacteriome, virome and the three metabolomic compartments into a single model. The analysis revealed that the first principal component of each dataset correlated with the others, indicating a coordinated, multi-system signature that distinguishes vegans from omnivores. Rather than isolated changes in one compartment, the vegan diet appears to orchestrate a synchronised restructuring across bacteria, viruses, microbial and host metabolites, and this integrated signature aligns with the vegans’ favourable clinical profile, which included lower glycaemia, lower atherogenic plasma lipids and reduced systemic inflammation as measured by high-sensitivity C-reactive protein.

Functional analysis of the microbial genomes added further texture. Vegans showed enrichment of bacterial genes encoding carbohydrate-active enzymes, the molecular machinery that microbes use to dismantle the complex polysaccharides abundant in plant cell walls. This is consistent with a community adapted to fermenting fibre, a process that yields short-chain fatty acids such as butyrate, propionate and acetate, compounds with well-documented anti-inflammatory and metabolic effects. The study also examined bacterial growth rates, estimated from sequencing coverage patterns across bacterial genomes, and found differences between the dietary groups, suggesting that the vegan gut is not just differently composed but differently dynamic.

The authors are careful about the limits of their design. This was a cross-sectional observational study of healthy volunteers, so it cannot prove that the vegan diet causes the microbial and metabolic differences; people who choose veganism may differ from omnivores in other relevant ways. The vegans had also followed their diet for varying lengths of time, and the team adjusted their analyses for age, sex and body mass index where appropriate. Nevertheless, the coherence of the findings, spanning diet records, microbial genomes, viral communities, protist prevalence, metabolites in three body compartments, cytokines and clinical chemistry, makes a compelling case that a well-constructed vegan diet is associated with a coordinated shift in gut ecology and metabolism that tracks with better cardiometabolic health. For anyone considering plant-based eating, the study carries a practical message: fibre is the engine of the vegan microbiome, and adequate plant protein may help recruit the most health-associated bacterial species of all.

Subject of Research: The association between vegan diet, gut microbiome composition, multi-compartment metabolome profiles and cardiometabolic health markers

Article Title: Vegans have distinct profiles of their gut microbiome and multi-compartment metabolome that align with markers of metabolic health

Article References: Thirion, F., Fan, Y., Hansen, T. H., Garcia-Perez, I., Lyu, L., Clos Garcia, M., Stankevic, E., Brix, S., Quinquis, B., Galleron, N., Levenez, F., Laiola, M., Maziers, N., Le Chatelier, E., Pons, N., Brignardello, J., Bello, A. T., Sarafian, M., Vestergaard, H., … Pedersen, O. (2026). Vegans have distinct profiles of their gut microbiome and multi-compartment metabolome that align with markers of metabolic health. Microbiome. https://doi.org/10.1186/s40168-026-02538-y

Image Credits: AI Generated

DOI: 10.1186/s40168-026-02538-y

Keywords: vegan diet, gut microbiome, metabolomics, shotgun sequencing, TMAO, Blastocystis, bacteriophages, short-chain fatty acids, cardiometabolic health, dietary fibre, p-cresyl sulphate, microbiome ranking