search

nextstrain / ncov

Nextstrain build for novel coronavirus SARS-CoV-2
https://nextstrain.org/ncov
MIT License
1.35k stars 403 forks source link

Nextstrain issue #584

Closed Leatherneck1 closed 3 years ago

Leatherneck1 commented 3 years ago

Current Behavior When using the online Nextstrain, none of the B.1.1.7 or P.1 strains (18 all told) show up when sorting either the Global or North America databases, and using Oregon as the filtering division. Also a search by specific ID (example USA/OR-OHSU-PHL00043/2021) does not show it in the database any longer. They used to mid February, not sure what changes broke the code.

Expected behavior For Oregon strains to show up on the "more aggressive" variants.

How to reproduce Steps to reproduce the current behavior: Visit https://nextstrain.org/ncov/north-america?dmin=2021-01-21&f_division=Oregon&l=radial&tl=author

  1. See no results showing for B1.1.7 or P.1. Others show, like the B1.1429/1427 aka "Cali" variants.

Possible solution I wish I knew

Your environment: if browsing Nextstrain online

Additional context Sample from Feb back when Oregon Division used to work: Nextstrain from Feb

jameshadfield commented 3 years ago

This is due to stochastic subsampling we use to reduce the entire GISAID dataset (700k genomes last time I checked) down to a manageable number of genomes for visualisation. Each time we rerun the analysis the included genomes will change.

There are now quite a few different analyses available on nextstrain (see https://nextstrain.org/sars-cov-2/ for a list). This SPHERES managed analysis focuses on Oregon and includes USA/OR-OHSU-PHL00043/2021.