Documentation for creating a pathogen repo

[joverlee521]

There are existing resources that we can point to as starting point:

• Alli's Epi Handbook
• Nextstrain VEME tutorial
• pathogen-repo-guide

However, most likely users will require additional guidance on general decisions as they set up their pathogen repo.
I imagine this to be a high level decision tree that can link out to other detailed docs.

Jotting down general questions and thoughts here that I think would fit into the decision tree:

• Do I need to create my own pathogen repo, use Nextclade, or fork an existing Nextstrain repo?
  • Is there an example for xyz pathogen? (https://docs.nextstrain.org/en/latest/guides/bioinformatics/examples.html#nextstrain-pathogens)
• Do I need to create an ingest workflow?
  • How do I create an ingest workflow? (Add tutorials for ingest pipelines #179)
• Do I need to create a Nextclade dataset?
  • How do I create a Nextclade dataset? (Nextclade dataset creation guide)
• How do I create a phylogenetic workflow?
  • Suggestion from Q3 planning mtg: Start with a plain language description of how we currently do things internally — find an existing similar pathogen (e.g., segmented, etc), and then adapt that. It could go here https://docs.nextstrain.org/en/latest/tutorials/creating-a-pathogen-repo/index.html and it would cover a lot of ground ahead of trying to do something more detailed
  • How do I choose a reference? (https://docs.nextstrain.org/en/latest/guides/bioinformatics/translate_ref.html)
  • How do I treat gaps? (https://docs.nextstrain.org/en/latest/guides/bioinformatics/missing-sequence-data.html)
  • How do I choose an aligner?
  • What's the minimum length requirement?
  • How do I decide which params to use for Augur commands?
  • Should I use BEAST or TreeTime?
    • If using augur refine/TreeTime:
      • how do I choose the --root?
      • when should I use --timetree (Fix Goofy date display lassa#50 (comment))?
      • do I need to specify a clock rate and how do I figure out clock rate?
  • How do I choose a subsampling scheme? (https://alliblk.github.io/genepi-book/sample-selection.html)
  • How do I implement my subsampling scheme? (e.g. Add docs for "tiered" subsampling #187)
  • How do I define lineages/clades for my pathogen?
  • How should I annotate AA muts (Docs for different methods to annotate AA muts #189)
  • Should I do a full genome or gene/CDS specific build?

[tsibley]

FWIW, I'd think to co-locate the docs for this topic within the pathogen-repo-guide repo itself. It seems like a nice central place to put them and they can stay in lock step with changes to the layout/organization in the guide.

[joverlee521]

Today's SAB meeting emphasized keeping simpler docs (e.g. zika-tutorial or zika-simple) to make starting out easier for new users.

@huddlej suggested:

We should include the new pathogen repo layout in the docs in a way that builds in the simpler Zika tutorial.

[joverlee521]

@jameshadfield had an idea on Slack to build the docs out from the quickstart Zika tutorials:

We could imaging a docs structure starting with the quickstart and then providing a number of extensions to this which you can choose as needed, for instance:

Entry point: Quickstart phylogenetics analysis (single Snakefile)

Then choose what functionality you would like to add:

• Ingest workflow
• Automation (using GitHub actions + pointers to use AWS if needed)
• Multiple builds (snakemake wildcards etc)
• Subsampling
• Add private data to existing data (either from ingest or from data.nextstrain.org)