Equine Uterine Microbiome — QIIME2 Workflow

DNA- and RNA-based 16S rRNA analysis of the low-biomass equine uterine microbiome

This repository contains the QIIME2 workflow developed to compare DNA- and RNA-based 16S rRNA V3–V4 amplicon sequencing approaches for the characterization of the low-biomass bacterial microbiome of the equine uterus.

The workflow accompanies the study:

Dyroff, A.I., López-Valiñas, Á., Magalhaes, H.B. et al. (2025).
Comparison of RNA- and DNA-based 16S amplicon sequencing to find the optimal approach for the analysis of the uterine microbiome.
Scientific Reports 15, 17037.
https://doi.org/10.1038/s41598-025-00969-5

Important: The additional files required to reproduce the analysis — including the trained taxonomic classifier, SILVA reference database, manifest, metadata and archived workflow resources — are available in the associated Zenodo repository:
https://doi.org/10.5281/zenodo.13847011

---

Workflow overview

Illumina paired-end reads
          │
          ▼
      QIIME2 import
          │
          ▼
 Cutadapt primer removal
          │
          ▼
     DADA2 denoising
          │
          ▼
 Feature table + ASVs
          │
          ▼
 Non-prokaryotic sequence removal
          │
          ▼
 Taxonomic classification
    SILVA / scikit-learn
          │
          ▼
 Chloroplast + mitochondrial removal
          │
          ▼
     ASV filtering
  abundance + prevalence
          │
          ├────────────────────┐
          ▼                    ▼
   Taxonomic tables      MAFFT alignment
   relative/absolute            │
                               ▼
                         Sequence masking
                               │
                               ▼
                         FastTree phylogeny
                               │
                               ▼
                       Midpoint-rooted tree
                               │
                               ▼
                       Diversity analyses
                     ┌─────────┴─────────┐
                     ▼                   ▼
              Alpha diversity       Beta diversity
          Shannon / Simpson / Chao1     Jaccard
                                         │
                                         ▼
                                 PERMANOVA / ADONIS

The workflow additionally generates taxonomic abundance tables across multiple taxonomic levels for downstream analyses.

---

Analysis strategy

Raw paired-end Illumina reads are imported into QIIME2 and primers are removed with Cutadapt before denoising and ASV inference with DADA2.

Representative sequences are filtered against the SILVA reference to remove non-prokaryotic sequences and subsequently classified using a trained SILVA / scikit-learn classifier. Mitochondrial and chloroplast sequences are excluded before downstream analysis.

Low-abundance ASVs are filtered using the strategy implemented in the original study: ASVs supported by at least 5 reads in at least 5 samples are retained, while otherwise excluded ASVs can be recovered when supported by at least 20 reads in one sample.

The filtered ASV dataset is subsequently used for phylogenetic reconstruction and diversity analyses. Representative sequences are aligned with MAFFT, hypervariable alignment positions are masked, and a phylogenetic tree is reconstructed using FastTree and midpoint rooting.

Alpha diversity analyses include Shannon, Simpson and Chao1 metrics. Beta diversity analysis includes Jaccard distances, group comparisons and PERMANOVA / ADONIS.

---

Software

The original analysis was performed using:

• QIIME2 2024.5
• macOS Ventura 13.4
• Apple M1 Pro
• osx-arm64

Additional tools used through QIIME2 or the command line include DADA2, Cutadapt, VSEARCH, MAFFT, FastTree and BIOM.

---

Data and resources

Raw sequencing data are publicly available through NCBI:

BioProject: PRJNA1165023

Additional resources required to reproduce the workflow are available on Zenodo:

Zenodo: 10.5281/zenodo.13847011

These resources include the trained classifier, SILVA reference sequences, manifest, metadata and archived analysis files and should be downloaded before running the workflow.

---

Running the workflow

1. Install QIIME2 2024.5.
2. Download the FASTQ files from BioProject PRJNA1165023.
3. Download the required reference and metadata files from the associated Zenodo repository.
4. Update the FASTQ paths in manifest.csv for your local system.
5. Run:

bash pipeline.sh

The complete analysis is provided in pipeline.sh, following the same sequence of processing, filtering, phylogenetic and diversity-analysis steps used in the associated study.

---

Citation

If you use this workflow, please cite the associated publication:

Dyroff, A.I., López-Valiñas, Á., Magalhaes, H.B. et al. (2025). Comparison of RNA- and DNA-based 16S amplicon sequencing to find the optimal approach for the analysis of the uterine microbiome. Scientific Reports 15, 17037. https://doi.org/10.1038/s41598-025-00969-5

The archived workflow and associated analysis resources are available from Zenodo:

López-Valiñas, Á. (2024). Unveiling the optimal approach for the analysis of the low biomass bacterial microbiome in the equine uterus: Comparison of RNA- versus DNA-based 16S rRNA V3-4 amplicon. Zenodo. https://doi.org/10.5281/zenodo.13847011

---

Author

Álvaro López-Valiñas, PhD
Bioinformatics · Microbiome · Genomics
ORCID: 0000-0002-7492-5108