Alisim: Only simulate InDels in existing sequence / Enable more complex dN/dS codon models

[casparbein]

Hi,

thank you for developing and maintaining AliSim. It works well for me, I just have two questions regarding codon-model simulations:

1. Would it be possible (and biologically reasonable) to only simulate InDels in an existing alignment? I have a simulated alignment without InDels (and the corresponding randomly generated tree), and I would like to just add InDels into this existing alignment without altering the underlying nucleotide sequence itself. I saw that when I use the "mimicking" option, the alignment itself will also change:

iqtree2 --alisim test_alignment \
-s input.noIndels.fa \
--indel 0.0008,0.005 \
--indel-size "GEO{2},GEO{2}" \
-t test_alignment.treefile \
-af fasta \
--seqtype CODON

2. The reason for the above question (why I use a different software to simulate the alignment) is that I need to simulate codon alignments with varying omega rates across sites and branches (say 70% of alignment positions of clade 1 with omega =1, 30% with omega = 4, 90% of alignment positions in clade 2 with omega = 1.5, 10% with omega = 10). As far as I could discern from the manual, currently, only tree-wide/clade-wide omega (dN/dS) values are supported, for example in the MG model. I know that I could create clade-specific omega values with [&model=MG{1.5}] or something like this, but is there a way to assign several omega values and their weights across specific clades/branches?

We don't have IQtree3 installed on our cluster yet, so I tested this with IQtree2, but I couldn't find anything about indel-only simulation and branch-site specific codon models in the current manual of IQtree3, either.

Thank you very much again for developing the software!
Cheers,
Bernhard