Reproducible Emu-Based Workflow for High-Fidelity Soil and Plant Microbiome Profiling on HPC Clusters

Accurate profiling of soil and root-associated bacterial communities is essential for understanding ecosystem functions and improving sustainable agricultural practices. Here, a comprehensive, modular workflow is presented for the analysis of full-length 16S rRNA gene amplicons generated with Oxford Nanopore long-read sequencing. The protocol integrates four standardized steps: (i) quality assessment and filtering of raw reads with NanoPlot and NanoFilt, (ii) removal of plant organelle contamination using a curated Viridiplantae Kraken2 database, (iii) species-level taxonomic assignment with Emu, and (iv) downstream ecological analyses, including rarefaction, diversity metrics, and functional inference. Leveraging high-performance computing resources, the workflow enables parallel processing of large datasets, rigorous contamination control, and reproducible execution across environments. The pipeline's efficiency is demonstrated on full-length 16S rRNA gene datasets from yellow pea rhizosphere and root samples, with high post-filter read retention and high-resolution community profiles. Automated SLURM scripts and detailed documentation are provided in a public GitHub repository (https://github.com/henrimdias/emu-microbiome-HPC; release v1.0.2, emu-pipeline-revised) and archived on Zenodo (DOI: 10.5281/zenodo.17764933). Key features • Implement rigorous quality control (QC) of raw 16S rRNA Nanopore reads and sequencing controls. • Remove plant organelle contamination with a curated Kraken2 database. • Perform high-resolution taxonomic assignment of full-length 16S rRNA reads using Emu. • Integrate downstream statistical analyses, including rarefaction, PERMANOVA, and DESeq2 differential abundance. • Conduct scalable microbiome diversity and functional analyses with FAPROTAX.