Augur Refine --root best clarification

Help and Getting Started
1

Hello,

I wanted to clarify the ‘best’ option when using augur refine --root

In the docs it says to run treetime -h for rooting method definitions: augur refine — Augur 34.3.0 documentation

--root

rooting mechanism (‘best’, ‘least-squares’, ‘min_dev’, ‘oldest’, ‘mid_point’) OR node to root by OR two nodes indicating a monophyletic group to root by. Run treetime -h for definitions of rooting methods.

Default: ['best']

When running treetime -h on the command line it only defines the following: "Valid choices are ‘min_dev’, ‘least-squares’, and ‘oldest’. "

In the treetime focs it lists best, least-squares beside each other.

Is “best” equivalent to least-squares? Alternatively, does it choose the best root out of all possible methods?

Thanks in advance!
Tara

2

Hello @tarabyte,

The ‘best’ value is an alias for ‘least-squares’, so they are equivalent.

This info was missing in older versions of the treetime help message, but should be in more recent versions, e.g. this is the help message I see for v0.12.1

--reroot REROOT [REROOT ...]
                        Reroot the tree using root-to-tip regression. Valid choices are 'min_dev', 'least-squares', 'oldest', and 'best' (alias for 'least-squares'). 'least-squares'
                        adjusts the root to minimize residuals of the root-to-tip vs sampling time regression, 'min_dev' minimizes variance of root-to-tip distances. 'least-squares' can be
                        combined with --covariation to account for shared ancestry. Alternatively, you can specify a node name or a list of node names to be used as outgroup or use
                        'oldest' to reroot to the oldest node. By default, TreeTime will reroot using 'best'. Use --keep-root to keep the current root.

Best,

Jover

3

Thank you very much, @joverlee!

I have a follow up question regarding the different root options and their impact on interpretation when comparing results between trees.

This is a bit connected to this post ( Measles root sequence ) where I received useful information regarding rooting and differences between the references and the root.

My use case of the tree is for looking at local samples with the global context, and how the trees change over time as new samples are added. If the root is variable depending on the samples in the dataset and their dates, I’m trying to determine the limitations or considerations I should be aware of, such as the impacts of the variable rooting on the divergence values.

I was wondering if it would be better to set a specific sequence as the root for comparable divergence values across different trees in this case. I’ve tested using a related virus (Rinderpest) as the root, but wasn’t sure about the impacts on clock rate.

Rinderpest virus NC_006296.2 as root:

image
image1920×1034 116 KB

TreeTime best root:

image
image1920×1044 178 KB

Regarding this section in the root vs reference documentation:

“In some sense, adding the reference sequence as a parent node to the tree makes the reference sequence “become” the root sequence, and that is terminology that is sometimes used in Nextstrain.”

If the reference sequence is added as the parent node, I wanted to confirm that is this purely related to updating the mutation coordinates relative to the reference positions and not necessarily a way to have the variable TreeTime rooting along with the stable divergence values (relative to the reference) between trees?

I hope that makes sense but please let me know if any parts are unclear.

Thank you for your time,

Tara