What are the suggested parameters for running IQ-TREE on a protein alignment for phylogeny?

Question

[O.rka]

I'm going through IQ-TREE's VERY VERY EXTENSIVE user manual and it's a bit intimidating. Are there any rules of thumb that would be a good default choice for most cases? I'm working with prokaryotes, viruses, and microeukaryotes.

Answer 1

[Mensur Dlakic]

Rather than going through the manual, I suggest reading their tutorials which explain certain cases.

• http://www.iqtree.org/doc/Advanced-Tutorial
• http://www.iqtree.org/doc/Substitution-Models
• http://www.iqtree.org/doc/Complex-Models
• http://www.iqtree.org/workshop/anu2018

I don't know if these are suggested parameters, but this is how I typically run IQ-TREE:

iqtree -s trimmed_alignment.phy -m TEST -nt 40 -vv -seed 1001 -pre trimmed_alignment.1st -t RANDOM

This is model finder and it will test common substitution models starting from a random tree. Obviously after -nt you put however many CPUs you have available, and use a random seed of your choice. Once the best model is found (look for the file trimmed_alignment.1st.iqtree from the names used above), it should be inserted into the next command and a first tree reconstruction can be done. This is 1000 fast bootstraps, which works fine for most cases.

iqtree -s trimmed_alignment.phy -m "BEST MODEL HERE" -nt 40 -vv -seed 1001 -t RANDOM -pre trimmed_alignment.2nd -bb 1000 -bnni -wbtl

Using the same names as above, the Newick tree would be in file trimmed_alignment.2nd.contree.

If you want to include mixture models into model search (takes longer), the first command becomes:

iqtree -s trimmed_alignment.phy -m MF -nt 40 -vv -seed 1001 -pre trimmed_alignment.1st -t RANDOM

That should be enough to get you going, but it is still a good idea to read through specific cases.