Make nextclade tree #28
Conversation
Creates a nextclade directory and copies relevant files from phylogenetic directory to nextclade directory. Subsequent commits will change these for the Nextclade Dataset tree workflow.
|
This is wonderful @kimandrews! Cleanly constructed in terms of the workflow and the result looks great in terms of consistency of MeV Genotypes and amount of data used. I think this will work well as a Nextclade reference tree. I just made a few changes to the Auspice config for the Nextclade tree in the above commit. |
There was a problem hiding this comment.
Thanks for working through this Kim! The overall workflow looks good to me, I only left some minor comments.
A couple high level thoughts (not blocking this PR):
- Will you be updating the workflow to create the full Nextclade dataset in the future? E.g. mpox's nextclade workflow packages the whole dataset.
- Since there are a lot of shared files between phylogenetic and nextclade workflows, we could consider using a shared directory to hold these files, so we don't need to duplicate the files across two directories.
nextclade/rules/export.smk
Outdated
| """ | ||
|
|
||
| rule export: | ||
| """Exporting data files for for auspice""" |
There was a problem hiding this comment.
| """Exporting data files for for auspice""" | |
| """Exporting data files for auspice""" |
Packaging up into a Nextclade dataset will be super useful. Would be nice to have this on the clades.nextstrain.org list.
I personally don't mind the code duplication in this particular case. Easier for me to think about these as separate things. But it's not a strong preference. |
There was a problem hiding this comment.
Nice work @kim, only a few minor requests from me. I didn't run this locally, but tested out the tree that it produces by using the use-nextclade-dataset branch which can also be loaded into the nextclade web UI and it's looking great.
* Add rule to align sequences to N450 reference using Nextclade, following edbefd5 * Filter to produce a sample set appropriate for a Nextclade Dataset tree
Remove construct_phylogeny.smk wildcards that were needed for the Phylogenetic workflow but are not needed for the Nextclade Dataset tree workflow.
* Use `augur ancestral` option `--root-sequence` to add mutations from the reference sequence to the root of the tree, which is required for Nextclade Dataset trees, as described here: https://github.com/nextstrain/nextclade_data/blob/577696e4ee0f0dff925d3911d27351a7395fab3d/docs/dataset-creation-guide.md * Use `augur clades` to assign genotypes to nodes * Remove annotate_phylogeny.smk wildcards that were needed for the Phylogenetic workflow but are not needed for the Nextclade Dataset tree workflow.
* Include `augur clades` genotype assignments in exported tree and in coloring options * Remove export.smk wildcards that were needed for the Phylogenetic workflow but are not needed for the Nextclade Dataset tree workflow.
This makes a few various updates to the Auspice config for the Nextclade tree including: 1. Making "clade_membership" default coloring. For this tree, we really care about MeV Genotype more than other colorings. 2. Call this "MeV Genotype (Nextstrain)". I think important to specify MeV Genotype to distinguish from the more common SNP-level genotype. 3. Clarify with "MeV Genotype (GenBank metadata)". The previous "(NCBI)" seemed like this could be a MeV Genotype call from an NCBI tool. 4. Order as Region then Country to follow ncov 5. Include clade_membership as filter option
kimandrews
force-pushed
the
make-nextclade-tree
branch
from
cabf98d
to
07bf737
Compare
|
Looks like this may be ready to merge? @jameshadfield: Could you confirm? |
I talked with @jameshadfield yesterday and we are good to merge. |
Description of proposed changes
This workflow creates a tree that can be used as part of a Nextclade dataset to assign genotypes to measles samples based on criteria outlined by the WHO
The tree created by this workflow includes N450 sequences for 28 reference strains defined by the WHO, along with other representative strains for each genotype that are obtained from the ingest output, using genotype assignments reported on NCBI. Genotypes are then assigned to each node with
augur clades, using clade-defining mutations indefaults/clades.tsv.Details regarding the workflow are described in a README.md
Related issue(s)
Checklist