How to remove third codon positions from a charset in iqtree?

Question

I need to build a phylogenetic tree using IQ-TREE, starting from a sequence alignment in CODON format of several invertebrate mitochondrial genes. These are my charsets:

nexus  
BEGIN SETS;

charset atp6.fasta = 1-651;  
charset atp8.fasta = 652-729;  
charset cob.fasta = 730-1749;  
charset cox1.fasta = 1750-3237;  
charset cox2.fasta = 3238-3858;  
charset cox3.fasta = 3859-4632;  
charset nad1.fasta = 4633-5484;  
charset nad2.fasta = 5485-6156;  
charset nad3.fasta = 6157-6459;  
charset nad4.fasta = 6460-7683;  
charset nad4L.fasta = 7684-7899;  
charset nad5.fasta = 7900-9294;  
charset nad6.fasta = 9295-9663;  
END;

I would need to remove the third bases, how should I write them? These are the commands with which I would launch IQ-TREE:

iqtree -s "fasta_file" -bb 1000 -alrt 1000 -m MFP+MERGE -spp "charset.nexus" -st CODON5

Please, any suggestions would be very helpful.

Answer 1

Yes, it's easy.

#nexus  
BEGIN SETS;

charset atp6.fasta = 1-651\3, 2-651\3;  
charset atp8.fasta = 652-729\3, 653-729\3;  
charset cob.fasta = 730-1749\3, 731-1749\3;  
# ... so on for every gene
END;

So for atp6 you will get 1,2,4,5,7,8,10,11 .... up to 651, and so on for each gene. Thus the residues at position 3, 6, 9, 12 are excluded.