Add `Nextclade` workflow for Norovirus genotyping

[j23414]

Context

There's a push to replace the Genomic Detective hardcoded files with a Nextclade dataset, as suggested in this comment and this comment. This change aims to improve automation and consistency in genotyping processes.

Creating a Nextclade dataset could require:

1. investigating clade-defining mutations or
2. creating a guide tree based on the Genome Detective results (ORF1_type and ORF2_type) and augur traits.

Current situation:

• A rough visual exploration of the samples shows that we have Genome Detective results for approximately 1,900 out of 51,047 norovirus samples with year information.
• Existing typing tools (such as CDC typing tool and RIVM typing tool) are web-only, limiting automation
• CDC provides a comprehensive list of reference sequences

A Nextclade dataset would greatly facilitate automated genotyping.

Potential next steps:

• Build a guide tree based on CDC reference sequences and identified clades
  * [ ] Scraped and cleaned into CDC_references (3).xls (estimated labor: 3+ hours) * due to recombination, genotype follows VP1 (capsid)
  * [ ] Pull capsid sequences and annotate header into norovirus_cdc_reference.fasta.txt (estimated labor: 15mins)
  * [ ] Optional: Pull reference sequences from Chhabra et al, 2019 Table 1 and duel nomenclature from Table 2

• Compare results against other tools

  • Compare against Genome Detective results for ORF1_type and ORF2_type
  • Compare against CDC typing tool
  • Compare against RIVM typing tool

• If results are consistent across tools, identify clade-defining mutations

  • Check for homoplasies which can result in inconsistent tree topologies
  • Mask any homoplasies that disagree with the collection date information (mutation similarities that arise from convergent or parallel evolution rather than from common ancestry)
  • Check for important indels compared to the chosen reference (evaluate the reference)
  • Check if the root needs to be different then the reference in the guide tree

• Develop Nextclade dataset files (reference sequence, pathogen.json , etc.)

• Implement the Nextclade workflow

• Implement rules that test the Nextclade dataset using example data

[xonq]

Hello @j23414 👋, I'm Zach with @theiagen. @AndrewLangvt and I are exploring options for making norovirus genotyping more accessible for state public health labs - building upon your team's awesome work and finalizing a Nextclade implementation is one option we are discussing. I'm curious if you're available to discuss more about this issue and how we might be able to help push implementation forward; please, feel free to contact me here or via email zachary.konkel@theiagen.com.

[j23414]

Just surfacing that I'm summarizing notes on a Norovirus Nextclade dataset on these slides

[j23414]

Sorry this is quite late, but summarizing key takeaways from meetings with Theiagen for documentation:

• Nextstrain and Theiagen folks met and discussed what a coordinated effort might look like on Feb 2nd, 2025
• On April 9th, Theiagen pivoted to work on rabies since it seemed simplier to tackle due to:
  • lack of recombination
  • existence of established software
  • RABV-GLUE using full-length genome sequences
• That collab resulted in Theiagen submitting a rabies nextclade dataset PR
• They ended up creating a "clade mutations extraction" script: extract-nextclade.py
• Email exchange documented on zohodesk