Combining DNA and morphological alignments with overlapping taxa (duplicated sequence name error)

[fernandasaluceste]

Hello,

I am trying to combine molecular and morphological datasets in IQ-TREE 3.

My datasets are:

Molecular alignment: molecular.phy

128 taxa

~3.2 million sites (UCE + multilocus DNA)

Morphological alignment: morpho.phy

90 taxa

196 discrete characters

Many taxa occur in both datasets, but the sampling is not identical.

Following the approach suggested in Discussion #213, I created a NEXUS partition file referencing the two alignments:

#nexus
begin sets;
charset molecular = molecular.phy: 1-3220459;
charset morphology = morpho.phy: 1-196;
end;

Then I ran:

iqtree3 -p partitions.nex -m MFP -B 1000 -alrt 1000 -T AUTO

IQ-TREE successfully loads the first alignment:

Loading 2 partitions...
Reading alignment file molecular.phy ...

and then the morphological alignment:

Reading alignment file morpho.phy ...

However it stops with the error:

ERROR: Duplicated sequence name farlowella_amazona26397
ERROR: Please rename sequences listed above!

This happens because several taxa occur in both alignments.

My question:

Is it possible in IQ-TREE to combine multiple alignment files that share taxon names (representing the same taxa across partitions), or should the datasets first be merged into a single supermatrix with missing data ("?")?

What is the recommended approach for combining molecular and morphological datasets with partially overlapping taxon sampling?

Thank you for your help.

[roblanf]

Hi, can you please share your two alignments so we can take a look?

Rob

[roblanf]

For what it's worth (and I'm not yet sure how much that is, but let's see...), Claude code thinks that this error is almost certainly because you have a duplicate name in your morpho alignment. If you can share your actual log file, that would also help.

Here's what Claude actually said:

Hi Fernanda,

IQ-TREE fully supports combined molecular + morphological partitioned analyses with overlapping taxa — your partition file approach is correct for this. When taxa appear in both files, IQ-TREE automatically builds a union of all taxa and fills in missing data where a taxon is absent from a partition.

I'd recommend explicitly specifying the sequence type for each partition to avoid auto-detection issues:

#nexus
begin sets;
charset molecular = molecular.phy: DNA, *;
charset morphology = morpho.phy: MORPH, *;
end;

Regarding the "Duplicated sequence name" error: this is triggered by a duplicate name within a single alignment file, not by shared taxa across partitions. Since the error appears after "Reading alignment file morpho.phy", the duplicate farlowella_amazona26397 is likely within morpho.phy itself. You can verify with:

grep -c "farlowella_amazona26397" morpho.phy

Once the duplicate is fixed, your command should work as-is:

iqtree3 -p partitions.nex -m MFP -B 1000 -alrt 1000 -T AUTO

IQ-TREE will automatically apply appropriate models to each partition (e.g., GTR-type for DNA, MK for morphology) via -m MFP.

[fernandasaluceste]

Hi Rob,

Thank you for your help.

You were correct — the issue was not caused by taxa appearing in both partitions. The problem was indeed a duplicated taxon name inside the morphological alignment itself. After checking the file, I found that farlowella_amazona26397 appeared twice in morpho.phy. Once I removed the duplicate entry, IQ-TREE was able to proceed normally.

Following the suggestion in Discussion #213, I am now running the analysis using a partition file that explicitly defines the data types:

#nexus
begin sets;
charset multilocus = molecular.phy:1-6449;
charset UCE = molecular.phy:6450-3220459;
charset morphology = morpho.phy: MORPH, *;
end;

The analysis now runs correctly and IQ-TREE loads all partitions:

Loading 3 partitions...
Subset 1 multilocus
Subset 2 UCE
Subset 3 morphology

It also correctly handles the partially overlapping taxon sampling by filling missing data for taxa absent in some partitions.

Thank you again for the clarification — this was very helpful.

Best regards,
Fernanda