Clustal Omega: Run an Alignment and Pick Output Format

Who this is for: a researcher with three or more related sequences who needs to produce a multiple sequence alignment with Clustal Omega, choose an output format the next tool can read, and know which setting to change when the result is not what was wanted.

Clustal Omega is a multiple sequence alignment program described in Sievers et al., Molecular Systems Biology 2011 (doi:10.1038/msb.2011.75). The authors report that it can align virtually any number of protein sequences quickly and deliver accurate alignments. On smaller sets its accuracy is similar to other high-quality aligners, and on larger sets it outperforms them in time and quality. It runs as a web service at the EMBL-EBI Job Dispatcher and as a command-line program, clustalo. This guide cross-walks the two, which the top-ranking pages tend to treat separately.

What Clustal Omega is used for in research

A multiple sequence alignment lines up three or more related sequences so that corresponding positions fall in the same column. Clustal Omega aligns protein sequences by default; the EBI service also accepts DNA and RNA through its sequence type setting (stype: protein, dna or rna). The 2011 paper highlights two further abilities: adding sequences to an existing alignment, and using information already in an alignment, including precomputed alignments held in public databases such as Pfam. To choose among aligners, see multiple sequence alignment tools compared. For the wider field, see what is bioinformatics.

Install or access

Online at EMBL-EBI

The EBI Job Dispatcher exposes Clustal Omega as the clustalo service. Its documented input rules are: three or more sequences; supported input formats GCG, FASTA, EMBL (nucleotide only), GenBank, PIR, NBRF, PHYLIP and UniProtKB/Swiss-Prot (protein only); and partially formatted sequences are not accepted. Source: EBI Job Dispatcher REST parameter documentation, read 2026-11-07.

Command line

The program is run as clustalo. The examples in the README distributed with the source code call it as ./clustalo from the directory that holds the program. Source: Clustal Omega README, read 2026-11-07, via a mirror of the upstream source on GitHub.

Step-by-step: align a FASTA file

1. Put three or more related sequences in one FASTA file, for example globin.fa (the README’s example file).
2. Run the default alignment: ./clustalo -i globin.fa.
3. Choose an output file and format with -o and --outfmt, for example Stockholm: ./clustalo -i globin.fa -o globin.sto --outfmt=st.
4. To write a ClustalW-style alignment and overwrite an existing file, add --force: ./clustalo -i globin.fa -o globin.aln --outfmt=clu --force.
5. To keep the by-products of the run, save the distance matrix and the guide tree: ./clustalo -i globin.fa --distmat-out=globin.mat --guidetree-out=globin.dnd --force.
6. To realign an existing alignment, strip its gaps first: ./clustalo -i globin.sto --dealign.
7. To refine with extra rounds, name the iteration count: ./clustalo -i globin.fa --iter=2. The README states that --iter must be given for iteration to happen.

On the web, the same decisions are made in the form: supply the sequences, set the sequence type, adjust the output options, and submit. Each form setting has a REST parameter name, matched to its command-line flag below.

Web setting to command-line flag crosswalk

EBI web/REST parameter	What it does (EBI documentation)	Command-line flag (README)
sequence	Three or more sequences, in a supported format	-i, --in or --infile (a file, or - for standard input)
outfmt	Alignment output format; default clustal_num	--outfmt
dealign	Remove existing gaps from the input; default no	--dealign
iterations	Combined guide-tree/HMM iterations, 0 to 5; default 0	--iterations or --iter=<n>
guidetreeout	Also return the guide tree; default yes	--guidetree-out
dismatout	Also return the distance matrix; only calculated if the mBed guide tree is switched off	--distmat-out

Some settings appear only in the EBI documentation, with no matching flag in the README option list: mbed (a faster, sampling-based guide tree for large sets; default yes), mbediteration, gtiterations (maximum guide-tree iterations within the combined iterations; default -1, range 0 to 5), hmmiterations and order. The README in turn lists flags the web parameter list does not include, such as --threads, --full, --full-iter, --guidetree-in and --hmm-in.

Output and how to read it

Choosing an output format

EBI outfmt value	Description (EBI)	README --outfmt value
clustal_num	ClustalW format with character counts (web default)	not listed
clustal	ClustalW format	clu or clustal
fa	Pearson/FASTA	a2m, fa or fasta
msf	MSF	msf
nexus	NEXUS	not listed
phylip	PHYLIP, interleaved	phy or phylip
selex	SELEX	selex
stockholm	Stockholm	st or stockholm
vienna	Vienna	vie or vienna

Pick the format named in the documentation of the program that will read the file next. If you need NEXUS, use the web service: the README option list does not include it.

Reading the alignment

Each sequence is one row and each column is an aligned position. Row order follows the order parameter: aligned (the default, the order determined by the alignment) or input. Gaps already present in an input alignment are kept unless you ask for them to be removed with dealign. The guide tree file (for example globin.dnd above) is the tree used to guide the alignment; to read tree files in general, see how to read a phylogenetic tree.

Troubleshooting

• The submission is rejected. Check that there are at least three sequences, that the format is one of those listed above, and that no record is only partly formatted.
• A downstream tool cannot read the file. Re-run with a different output format from the table.
• The command will not write an output file. The README examples add --force when overwriting a file.
• No distance matrix came back on the web. It is only calculated when the mBed guide tree is switched off.
• Existing gaps are carried into a new run. Use dealign (web) or --dealign (command line).
• No refinement happened. Iterations default to 0 on the web, and on the command line --iter must be given.

Free alternatives

Other aligners are compared side by side in multiple sequence alignment tools compared. For protein records to feed into any aligner, see UniProt, the protein sequence and functional annotation knowledgebase.

FAQ

How many sequences does Clustal Omega need?

Three or more. The EBI service documents “three or more sequences” as the input requirement.

Can it align DNA and RNA as well as protein?

The EBI service offers protein (default), DNA and RNA as sequence types. The 2011 paper itself describes protein alignment.

Can I add sequences to an existing alignment?

The paper describes features for adding sequences to an existing alignment and for using information in it, including precomputed alignments in databases such as Pfam. The README option list includes --hmm-in and --guidetree-in.

Which output format is the default?

On the EBI service, clustal_num: ClustalW format with character counts.

Does it save the guide tree?

The EBI service returns the guide tree by default (guidetreeout), and the command line saves it with --guidetree-out.

Sources

• Sievers F, et al. Fast, scalable generation of high-quality protein multiple sequence alignments using Clustal Omega. Mol Syst Biol 2011;7:539. doi:10.1038/msb.2011.75. PMID 21988835. PubMed record, read 2026-11-07.
• EMBL-EBI Job Dispatcher, Clustal Omega REST parameter list and parameter details, read 2026-11-07.
• Clustal Omega README (mirror of the upstream source), raw README, read 2026-11-07.