Tamás Járay

Diet changes the fecal metabolome quickly, but it is less clear how the metabolome recovers once the original diet is restored. We followed nine Pumi dogs through an owner-managed switch from dry food to a raw (BARF) diet and back, and profiled 72 fecal samples by untargeted UPLC–MS. Diet phase accounted for roughly half of the dog-centered variation in both ionization modes. At the first sampling point after the switch to raw food, more than 13,000 LC–MS features had changed, and the return to dry food produced a response of similar size in the opposite direction: of the features that changed at both transitions, more than 99% reversed. By the last sampling point, no ESI+ features and only 13 ESI− features met the significance and effect-size thresholds for a difference from the second dry-food baseline. The overall BARF-associated pattern persisted when individual dogs were left out, and the displayed candidate associations remained significant in pedigree-adjusted models, although individual feature effects depended on normalization. Putative indole/tryptophan, bile-acid, and aromatic amino-acid metabolites showed different response and recovery patterns. In this cohort, the fecal metabolome changed rapidly during the dietary transition and returned largely toward baseline after dry food was resumed, with differences in recovery among dogs.

Long-read metagenomic sequencing can improve genome recovery from complex gut microbial communities, yet directly reusable canine gut genome resources remain limited. Here we describe DogMAG, a canine gut metagenome resource based on dog-wise long-read and hybrid assemblies generated by grouping sequencing libraries according to canonical dog identity before assembly. The final dataset comprises 41 assemblies linked to 277 FASTQ records, including 30 Flye long-read-only and 11 OPERA-MS hybrid assemblies. A single integrated BASALT workflow produced 11,276 selected bin/version records, followed by explicit quality-based re-selection of 3,418 medium-quality-or-better metagenome-assembled genome candidates. External dRep dereplication yielded 792 strain-like representatives at 99% average nucleotide identity and 135 species/SGB-like representatives at 95%. GTDB-Tk classified all 792 representatives as Bacteria. Viral screening identified 22,068 geNomad predictions, of which 3,374 Complete, High-quality or Medium-quality viral/proviral candidate rows passed CheckV filtering with contamination ≤10%. DogMAG provides assemblies, genome and viral candidate sequences, metadata, provenance tables and workflow scripts for reuse, benchmarking and reanalysis.

Human spaceflight may perturb the gut microbiota, but densely sampled short missions remain poorly characterized. We profiled 27 phase-matched fecal samples from two astronauts during an 18-day International Space Station mission and one ground-based participant following the same daily schedule using Oxford Nanopore full-length 16S sequencing. Participant identity dominated genus-level Bray- Curtis variation (R2 = 0.489, p < 0.001). In astronaut-only community analyses, mission phase explained 23.4% of genus-level (p = 0.035) and 22.3% of species-level (p = 0.021) variation. Astronauts showed greater displacement from personal baselines than B1 (0.331 versus 0.171) and 1.58-fold higher volatility. Astronaut-only taxon models identified 2 of 81 genera and 5 of 139 species; Collinsella increased from quarantine to orbit (coefficient = 2.586, q = 0.037). Thus, the short-duration spaceflight interval was accompanied by individualized, temporally localized community and taxon shifts rather than uniform microbiota restructuring.

Early-life gut microbiome maturation remains poorly resolved in dogs. We analyzed 146 fecal samples from 19 Pumi puppies across four dietary stages using paired shotgun metagenomics and Oxford Nanopore full-length 16S sequencing. Bacterial communities shifted from Escherichia -rich facultative assemblages during breastfeeding to anaerobe-rich communities after the transition to solid food, with the strongest compositional change occurring during dietary diversification. Functional profiles changed most strongly at the start of complementary feeding, while viral profiles showed increasing phage richness and stage-dependent predicted phage–host associations. Despite this shared dietary trajectory, host identity explained a substantial fraction of microbiome variation and produced persistent dog-specific abundance patterns. Long-read 16S profiles recovered the main developmental signal despite platform-dependent taxonomic bias, and 16S-based functional inference captured pathway-level contrasts concordant with matched metagenomes. These results identify early-life canine gut microbiome maturation as an age-diet-associated but host-constrained ecological process involving coordinated bacterial, viral, and functional restructuring.

The canine gut microbiome serves as a key model for veterinary and human health research, but inconsistent findings arise due to methodological variations. This study presents a three-part dataset to clarify how DNA extraction, primer selection, and sequencing platforms influence microbial profiling. First, we performed ultra-deep sequencing of a single dog fecal sample using five DNA isolation kits, multiple library protocols, and four sequencing platforms (Illumina MiSeq/NovaSeq, ONT MinION, PacBio Sequel IIe), enabling direct comparisons of 16S rRNA and shotgun sequencing techniques. Second, we analyzed 40 fecal samples from eight co-housed dogs using Zymo High-Molecular-Weight (ZHMW) and Zymo MagBead (ZMB) extraction kits to assess longitudinal extraction-kit effects. Third, we evaluated three 16S primer systems (standard ONT, PacBio, and modified ONT with degenerate bases) using synthetic mock communities and human/canine fecal samples to quantify primer biases. By integrating synthetic and biological replicates, this dataset provides a standardized resource for benchmarking bioinformatics pipelines and improving cross-study comparability. The study generated 75.31GB of new sequencing data: 43.451GB from ZHMW-ZMB comparisons, 22.611GB for primer evaluations, and 9.191GB from the single-sample analysis. Combined with 31.51GB of prior data, the total dataset exceeds 1061GB, including all analytical outputs. These resources enhance methodological transparency and accuracy in canine gut microbiome research across diverse laboratory workflows.

A critical issue in microbiome research is the selection of reliable laboratory and bioinformatics pipelines. In the absence of generally accepted technical benchmarks and evaluation standards, comparing data generated by different studies becomes challenging. In this work, we carried out the most comprehensive study to date on this topic. We encompassed every stage of processing, from DNA extraction to computational assessment. We adopted four procedures for DNA purification, six for library construction, three for sequencing, and five for bioinformatics. Additionally, we used datasets published by others to corroborate our results. We introduced a software tool that distinctively delivers consistent results, irrespective of sample or dataset origins. This study underscores the importance of methodological optimization at the outset of research projects to ensure the reliability of results and their comparability with findings from other studies. Additionally, this study provides an optimized robust pipeline for gut microbiome analysis.