Dealing with very large trees#
Loading very large trees in Taxonium on some platforms may sometimes require a few tricks. Here we summarise information on this. We are continuing to try to understand the limitations on different platforms.
Situations that are straightforward#
Here we summarise some situations of large trees that load without requring any tweaks:
Loading cov2tree on almost any platform, which uses the Taxonium backend so that you can explore a tree of >5M sequences without having to download them locally
Locally loading the public SARS-CoV-2 tree of 5M nodes on ~any platform with at least a few GB of RAM. This will also apply to trees of similar size with similar amount of metadata
Loading a tree of >11M SARS-CoV-2 sequences (including sequences from GISAID) in Chrome on Windows
Situations that may be more complex#
We have observed a memory limitation somewhere between 5M and 11M sequences (dependant also on amount of metadata and mutations) on Chrome for MacOS. This appears to be due to a 2GB memory limit on web worker processes specific to this platform. Using Firefox appears to avoid this issue. We recommend the use of Firefox if using macOS to view very large trees locally.
In some situations you might want to deploy your own Taxonium backend to avoid these issues.