Align sequences and build a phylogenetic tree

What this involves

Inferring an accurate phylogenetic tree begins with strict sequence curation. Uncleaned sequence inputs—such as residual gaps, internal stop codons, or unhandled soft-masking—corrupt alignment columns and downstream tree topologies. Before computing alignments, sequences must be validated for length, alphabet composition, and ambiguity codes, while homologous sets should be curated to avoid severe length mismatches.

Multiple sequence alignment is performed using dedicated aligners such as MAFFT or MUSCLE, or clustered first with CD-HIT to reduce redundancy. Tree inference follows from these alignments using maximum-likelihood methods. IQ-TREE provides substitution model selection and branch support analysis, while FastTree delivers approximately-maximum-likelihood trees suitable for larger sequence collections.

The research agent executes these command-line tools directly on its own environment, handling tool configuration, file input and output, and parameter validation. Nothing needs to be installed locally; the agent reads tool outputs back and reports the exact commands, metrics, and parameters used.

How it goes

1. 1Clean and characterise input sequencesInput strings are stripped of FASTA headers, non-alphabetic characters, and gaps, then uppercased to preserve soft-masked residues. The agent checks alphabet contents, noting stop codons or ambiguity codes (X, B, Z, J) that affect sequence interpretation.
2. 2Reduce redundancy with CD-HITWhen sequence collections are large or heavily biased, CD-HIT clusters protein or nucleotide sequences by identity to produce a representative, non-redundant set for downstream alignment.
3. 3Build the multiple sequence alignmentThe agent runs MAFFT or MUSCLE to compute a multiple sequence alignment from the cleaned fasta sequences, writing aligned output files to the workspace.
4. 4Infer the phylogenetic treeUsing the aligned sequences, the agent infers a phylogenetic tree using IQ-TREE for full maximum-likelihood inference with model selection and branch support, or FastTree for approximately-maximum-likelihood trees on large alignments.

Start it here

Ask in plain language. The agent works on a Linux machine of its own — it installs and runs what the job needs, keeps every file, and reads the results back with the numbers explained. Nothing to set up, and a long run continues after you close the page.

What to watch for

• Deleting lowercase characters instead of uppercasing them alters sequence lengths and shifts every downstream residue index.
• Percent identity without aligned length and query coverage is meaningless; high identity over a short window represents a local motif rather than homology.
• E-values scale directly with database size, so compare bit scores rather than E-values across different searches.
• Between 20% and 35% pairwise identity lies the twilight zone, where structural fold may be conserved but function often is not.
• An internal asterisk in a translated sequence indicates a frameshift or pseudogene that must be corrected before proceeding with alignment.
• Default or zero gap penalties shred alignments and produce severely inflated identity estimates.

What it runs

Questions