Fix cattle-outbreak strain selection #133
Labels
bug
Something isn't working
Comments
|
The automated run produced (essentially) this tree over the weekend. I've disabled the GitHub action and re-uploaded the https://next.nextstrain.org/avian-flu/h5n1-cattle-outbreak/genome@2025-02-20 dataset as the latest. |
Merged
jameshadfield
added a commit
that referenced
this issue
Feb 24, 2025
The previous approach relied on broad filtering -- minimum date of 2024, region of North America -- a hardcoded exclude list and a clock filter. As the diversity of sequences increased the clock-filter became less effective and ultimately dropped all the desired strains. See <#133> for more. We now use GenoFLU constellations and can relax the date and region filters accordingly. This relaxation didn't result in any non-North-American samples included but did add one B3.13 genome from 2023: 'A/Goose/USA/23-038138-001-original/2023'. These changes to filtering will also apply to the D1.1 builds, but testing indicates no changes. The segment-level approach is not addressed here, but could be similarly adjusted to use the GenoFLU matching on the segment level. Specifically, <https://github.com/nextstrain/avian-flu/pull/138/files> adds the segment-level annotations and the expanded constellation is: B3.13 = PA:ea1, HA:ea1, PB1:am4, MP:ea1, NA:ea1, PB2:am2.2, NP:am8, NS:am1.1 Closes #133
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
In the process of fixing #131 (via #132) the genome tree was completely wrong due to filtering out all the strains we actually want! (Screenshot below). The strains were dropped by the clock filter in
augur refine. Now that GenoFLU metadata is (or soon will be) available via #126 we should try restricting toB3.13.The text was updated successfully, but these errors were encountered: