nextstrain / forecasts-ncov

SARS-CoV-2 variant growth rates and frequency forecasts
https://nextstrain.org/sars-cov-2/forecasts/
7 stars 2 forks source link

Add clades 24H, 24I #112

Open corneliusroemer opened 1 week ago

corneliusroemer commented 1 week ago

I don't know how to update colors

joverlee521 commented 1 week ago

I'll take more detailed look at this tomorrow.

Noting for myself that the forecast page currently does not have the new clades:

Screenshot 2024-11-18 at 5 12 39 PM

This is likely because the sequences counts do not reach the clade_min_seq config https://github.com/nextstrain/forecasts-ncov/blob/31a6163d76254ea53cd8def4e19ff6b5f8c77176/config/config.yaml#L23

We can force include the new clades with something like (untested):

diff --git a/config/config.yaml b/config/config.yaml
index f0a2307..787ce8e 100644
--- a/config/config.yaml
+++ b/config/config.yaml
@@ -22,6 +22,7 @@ prepare_data:
         prune_seq_days: 12
         clade_min_seq: 2000
         clade_min_seq_days: 150
+        force_include_clades: "24H=24H 24I=24I"
     pango_lineages:
       global:
         included_days: 150
corneliusroemer commented 1 week ago

I think we should lower thresholds in general. There are many interesting countries with relevant sequencing activity that are excluded. I did a test run with lower thresholds across the board and it looks good. Will share.

Sequencing activity keeps dropping, so without dropping thresholds the forecasts become increasingly irrelevant.

Same for pango lineages. LP.8 has around 140 sequences now. It's growing fast, it's been noteworthy for multiple weeks but will show only in a few weeks with current thresholds demanding 350 sequences.

I can make a proposal tomorrow.