Using AliSim to generate synthetic sequences that are similar to input

[David Cowan]

Hi, I like the AliSim and IQTree tools; however, I haven't been able to do what I want it to do. I am trying to use AliSim to generate sequences that similar to the input sequences.

So far I have used '‘in\iqtree2 --alisim data/test2 -s fivePureSubtypesMax2withAncestor.fas --single-output -af fasta --num-alignments 2 --root-seq fivePureSubtypesMax2withAncestor.fas,SIVcpzMB66' to generate sequences. Then, I update the output fasta so that each batch can identified by add 'a' or 'b' to end of name. Additionally, I add my original input fasta sequence to this file. I use that file to generate a tree with ‘bin\iqtree2 -s data/test2.fa -redo’.

 When looking at the tree (exampleTree.png), I see that sequences from each batch are more similar to that batch than to those of the same subtype. However, I am interested in generating synthetic sequences that don't just match the tree, but are more similar to the same subtype/taxon as in expected.png.

Is this possible with AliSim and what parameters should I test?

Thanks,

David

[David Cowan]

Just following up on my question. Let me know if its not clear what I am trying to do. Essentially if given a tree with node (A,B,C,D,E...),  I am currently able to generate sequences that are closer to other generated sequences in that batch than to it original sequence. So A' is closer to B',C',D',E'... than it is to A; B' is closer to A', C', D', E' than to B.... My goal is to generate sequences so that each generated sequence is closest distance wise to its original sequence node. For example, A' will be closest to A, B' will be closest to B, etc... 

Is my goal possible with AliSim and IQTree? If so, what parameters should I try?

Thanks,
David