Use Pathoplexus

[victorlin]

Pathoplexus provides API endpoints for metadata and sequences of various pathogens.

This issue tracks its usage across our pathogen repos. The list below is the intersection of all Nextstrain pathogens and Pathoplexus organisms (as of 2026-03-17).

• rsv (over several PRs)
• ebola
  • Note: display name, location data, and temporal data were insufficient and are enriched by pulling from other sources.
• mpox
• WNV
• hmpv
• measles
• dengue
• yellow-fever

Open questions / things to consider when using it in a pathogen repo:

• Certain patterns should be followed, such as inclusion of data use terms and links in the Auspice tip info modals.
• Should the restricted sequences be included in our analyses?
• Is the Pathoplexus metadata sufficient as-is? Specifically:
  • display name
  • location data, e.g. ingest: division includes detailed location rsv#99
  • temporal data
• What downloads should be allowed after Allow JSONs to control downloading ability auspice#2047 has been released?

[joverlee521]

Should the restricted sequences be included in our analyses?

Yes! I think the availability of the restricted sequences is one of the major wins of using PPX data. We need to make sure that we clearly indicate their use terms in the curated metadata on S3 and in our builds in Auspice.

I haven't outlined this anywhere, but I'm in the process of updating RSV to follow a new pattern with the multiple inputs for the phylogenetic workflow (nextstrain/rsv#97):

Ingest

• after curation, split metadata/sequences into OPEN vs RESTRICTED data
• upload them as separate files on S3 so that the RESTRICTED data is clearly indicated in the file name

  terms	S3
  OPEN	files/workflows/<pathogen>/metadata.tsv.zst
  	files/workflows/<pathogen>/sequences.fasta.zst
  RESTRICTED	files/workflows/<pathogen>/metadata_restricted.tsv.zst
  	files/workflows/<pathogen>/sequences_restricted.fasta.zst

Phylogenetic

• use multiple inputs to include both OPEN and RESTRICTED data for our analyses

  inputs:
      - name: ppx_open
        metadata: https://data.nextstrain.org/files/workflows/<pathogen>/metadata.tsv.zst
        sequences: https://data.nextstrain.org/files/workflows/<pathogen>/sequences.fasta.zst
      - name: ppx_restricted
        metadata: https://data.nextstrain.org/files/workflows/<pathogen>/metadata_restricted.tsv.zst
        sequences: https://data.nextstrain.org/files/workflows/<pathogen>/sequences_restricted.fasta.zst

[joverlee521]

I'm in the process of updating RSV to follow a new pattern with the multiple inputs for the phylogenetic workflow

See PRs in RSV

• phylogenetic: Add standardized multiple inputs rsv#108
• Separate files for RESTRICTED sequences rsv#112

[victorlin]

The implementation of adding additional URL columns now varies significantly across repos.

• rsv: Uses a custom script in rule curate to add
  • PPX_accession__url
  • INSDC_accession__url
• ebola: Uses csvtk in rule add_accession_urls to add
  • PPX_accession__url
  • INSDC_accession__url
• mpox: Uses a custom script in rule curate to add
  • PPX_accession__url
  • PPX_accession_version__url
  • INSDC_accession_version__url
  • Pathoplexus_group__url
  • submission_database
• others are mostly duplicates of the custom scripts above, see GitHub query

It would be nice to centralize implementation. Maybe a new augur curate command to replace the custom scripts would be appropriate?