Why is nuc:7851T (AY.4 defining) such a mess on the Europe tree?

[corneliusroemer]

nuc:7851T is not masked. What's going on here? Why is it not all on one branch? Is this a broader iqtree problem just showing up drastically like this or something else?

Build for example: https://nextstrain.org/groups/neherlab/ncov/europe/2021-11-02?c=pango_usher&gt=nuc.7851T&m=div

[corneliusroemer]

So strange, why is it suddenly a problem and never before?
https://nextstrain.org/groups/neherlab/ncov/europe/2021-11-02?c=gt-nuc_7851,17040&m=div

[rneher]

I think this is a tree building problem. Masking has made the polytomy bigger and iq-tree struggles.

[corneliusroemer]

Is there something we can do to help iq-tree? Like increase the number of iterations, seeds, etc? It only runs for ~20min which in the grand scheme of things is very short and could easily be extended without incurring any significant costs.