Nextstrain build for Orthohantavirus

This repository contains the Nextstrain build for Orthohantavirus (hantavirus). It fetches genomic data from NCBI, classifies each sequence by species and genome segment, curates the metadata, builds per-species/segment phylogenies, and exports Auspice datasets for interactive visualization.

Hantaviruses have a tri-segmented negative-sense RNA genome — the L (large), M (medium), and S (small) segments — so builds are produced independently for each species/segment combination.

Repository organization

The build is split into two independent Snakemake workflows, following the standard Nextstrain pathogen-repo layout:

Directory	Purpose
ingest/	Fetch sequences from NCBI, classify them by species/segment, and curate metadata.
phylogeny/	Align sequences, build and refine trees, annotate, and export Auspice JSONs.
shared/	Per-species/segment RefSeq GenBank files used as alignment and annotation references.

The two workflows are connected only through the filesystem: phylogeny/ reads the curated output written to ingest/results/. There is no top-level Snakefile — each stage is run separately.

Installation

The workflows depend on Augur and a number of bioinformatics tools. The Nextstrain nextstrain-base conda package bundles most of them (augur, snakemake, the NCBI Datasets CLI, csvtk, iqtree, plus biopython/pandas/numpy), but not diamond, which the ingest workflow needs for sequence classification. Create an environment with both in one command:

mamba create -n nextstrain -c conda-forge -c bioconda -c nextstrain nextstrain-base diamond

Then activate it before running either workflow:

conda activate nextstrain

Alternatively, use the Nextstrain CLI with a managed runtime, ensuring diamond is additionally available on the PATH.

Usage

Each workflow uses workdir: workflow.current_basedir, so run it from within its own directory (or point Snakemake at its snakefile); all paths inside the rules are relative to that directory.

1. Ingest

cd ingest
snakemake --cores 4

This downloads the NCBI virus dataset for the configured taxon, assigns each sequence a species and segment by aligning it against RefSeq proteins with DIAMOND, filters by percent identity, and runs the augur curate pipeline. Curated output is written to:

ingest/results/{species}/{segment}/metadata_curated.tsv
ingest/results/{species}/{segment}/sequences_curated.fasta

2. Phylogeny

cd phylogeny
snakemake --cores 4

This filters and aligns the curated sequences, builds a tree with IQ-TREE, refines it (midpoint rooting by default), reconstructs ancestral nucleotide and amino-acid mutations, assigns colors, and exports one Auspice dataset per species/segment:

phylogeny/auspice/{species}_{segment}.json

Use snakemake -n (dry run) in either directory to preview the DAG before running.

3. Visualize

auspice view --datasetDir phylogeny/auspice

Configuration

Most behavior is controlled through the defaults/config.yaml file in each workflow, without editing rule logic:

• ingest/defaults/config.yaml — NCBI taxon ID, metadata field mapping, and the augur curate settings (strain-name regex, date formats, titlecasing, geolocation rules).
• phylogeny/defaults/config.yaml — per-species/segment sequence length bounds, rerooting strategy, and the metadata columns exported to Auspice.

The set of species and segments to build is defined as Python lists at the top of each workflow's snakefile (ingest/snakefile, phylogeny/snakefile) — edit those to add or remove a build. Other tunable inputs include ingest/defaults/rename_species.csv (species name normalization), ingest/defaults/special_cases.csv (manual classification overrides), and phylogeny/defaults/exclude.txt (strains to drop).

References

Corresponding RefSeq references for each species/segment live in shared/ as {species}_{segment}_refseq.gb. They serve as both the alignment reference and the gene-annotation source for amino-acid translation.