In the phylogenetic workflow, we initially listed strains slated to be dropped from the build (due to excessive divergence or misclassification) in the config/dropped_strain.txt file, identified by their strain names. Subsequently, we transitioned to using "accession" numbers to identify strains after merging the ingest pipeline (and using ncbi-datasets api) https://github.com/nextstrain/dengue/commit/8ab810f10cc64e45a9c2baf1bf6df9d6561578bc. However, updating the corresponding dropped strain list was missed resulting in these strains not being appropriately dropped from the build.
This commit addresses this issue by adding accession numbers to ensure proper dropping of these records.
Steps to find accessions:
For strain whose names exactly match those in the faunametadata.tsv file, the associated accession numbers are used.
A visual check is performed for the remaining strains, identifying similarly named strains
(e.g. DAK_Ar_510 is probably a shorter name for DENV2/COTE_D_IVOIRE/DAKAR510/1980)
Strains are cross-referenced against the ingestmetadata.tsv file for those collected in the same year and country, considering potential differences in underscores or hyphens separating the strain name (e.g. DENV1/VIETNAM/BIDV992/2006 equivalent to DENV-1/VN/BID-V992/2006).
Perform a rough search of dengue academic papers to find a similarly named strain, and reference to an accession number. (e.g. A search for DENV/SPAIN/EEB17/2009 led to https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3149010/ which led to accession number JF260983)
Description of proposed changes
In the phylogenetic workflow, we initially listed strains slated to be dropped from the build (due to excessive divergence or misclassification) in the config/dropped_strain.txt file, identified by their strain names. Subsequently, we transitioned to using "accession" numbers to identify strains after merging the ingest pipeline (and using ncbi-datasets api) https://github.com/nextstrain/dengue/commit/8ab810f10cc64e45a9c2baf1bf6df9d6561578bc. However, updating the corresponding dropped strain list was missed resulting in these strains not being appropriately dropped from the build.
This commit addresses this issue by adding accession numbers to ensure proper dropping of these records.
Steps to find accessions:
metadata.tsvfile, the associated accession numbers are used.DAK_Ar_510is probably a shorter name forDENV2/COTE_D_IVOIRE/DAKAR510/1980)metadata.tsvfile for those collected in the same year and country, considering potential differences in underscores or hyphens separating the strain name (e.g.DENV1/VIETNAM/BIDV992/2006equivalent toDENV-1/VN/BID-V992/2006).DENV/SPAIN/EEB17/2009led to https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3149010/ which led to accession numberJF260983)Related issue(s)
Checklist
Post Merge rebase and update checklist