Alignment and phylogenetics · Software
IQ-TREE
Infers maximum-likelihood phylogenetic trees with automated model selection and ultrafast branch support.
What IQ-TREE is
IQ-TREE is an open-source phylogenomic software suite designed to infer evolutionary trees using maximum likelihood. Created as the successor to IQPNNI and TREE-PUZZLE, it addresses the computational challenges brought on by the rapid growth of phylogenomic data, providing scalable tree searches alongside complex models of sequence evolution.
The software provides an end-to-end phylogenomic toolkit, integrating substitution model selection through ModelFinder, branch support assessment using an ultrafast bootstrap approximation (UFBoot), and synthetic data generation via the AliSim sequence simulator. It can harness multicore processors and distributed parallel computing environments, and automatically writes checkpoints to allow users to resume interrupted computations.
IQ-TREE supports an extensive range of evolutionary scenarios, handling DNA, amino acid, codon, binary, and morphological datasets. It features advanced frameworks including partition models, site- and tree-mixture models, polymorphism-aware approaches (PoMo), and large-scale pathogen phylogenetic methods.
How it works
Phylogenetic inference in IQ-TREE relies on a fast, stochastic maximum-likelihood tree search algorithm that optimises tree topology and branch lengths. ModelFinder assesses substitution models across datasets and can evaluate optimal partitioning schemes. For branch support, the ultrafast bootstrap (UFBoot) provides an approximation that operates substantially faster than standard non-parametric bootstrapping while reducing support bias. Single-branch tests, such as SH-aLRT and aBayes, as well as tree topology tests like the approximately unbiased (AU) test, are also available.
The platform incorporates complex evolutionary models, including partition models with linked or unlinked branch lengths across loci, empirical and user-defined mixture models (including MixtureFinder and GTRpmix), heterotachy models (GHOST), and tree mixture models (MAST) that capture multiple evolutionary histories across sites. For massive genomic datasets, specialized modules such as CMAPLE allow phylogenetic inference on millions of pathogen sequences, while QMaker can estimate empirical protein substitution matrices.
Key facts
- Underlying method
- Maximum-likelihood phylogenetic inference via a stochastic search algorithm
- Supported data types
- DNA, protein, codon, binary, and morphological data
- Accepted input formats
- PHYLIP, FASTA, Nexus, Clustal, and MSF sequence alignments
- Output formats
- Self-readable report files (.iqtree) and NEWICK tree files (.treefile)
- Parallel computing
- Multicore and distributed parallel processing support
- Fault tolerance
- Automatic checkpointing to resume interrupted analyses
- Supported operating systems
- 64-bit Windows, macOS (Universal, Intel, ARM), and Linux (Universal, Intel, ARM)
What you can use it for
- Phylogenomic Tree Inference and Model Selection
- Reconstruct maximum-likelihood phylogenetic species or gene trees from multi-locus or genome-wide alignments while automatically determining the best-fit substitution and partitioning schemes.
- Branch Reliability Assessment
- Evaluate confidence in phylogenetic branch splits rapidly across large alignments using ultrafast bootstrap approximation (UFBoot), SH-aLRT, or aBayes branch tests.
- Large-Scale Pathogen Surveillance
- Reconstruct evolutionary histories from massive outbreak datasets comprising millions of viral or bacterial genomes, such as SARS-CoV-2, using the CMAPLE module.
- Alignment Simulation
- Generate synthetic sequence alignments under standard or complex evolutionary models using the built-in AliSim simulator.
- Phylogenomic Discordance and Mixture Modelling
- Evaluate gene and site concordance factors (gCF and sCF) or apply mixture models (such as MAST or GHOST) to account for heterotachy and varying evolutionary histories across sites.
Run it here
IQ-TREE is preinstalled on the machine behind every research conversation. Ask the agent in plain language, or give it the exact command — it runs it, reads the output and explains it.
Try asking
“Infer a maximum-likelihood tree with ultrafast bootstrap from this alignment using IQ-TREE.”
Strengths and limits
Good at
- Reconstructs maximum-likelihood phylogenies with comparable likelihoods to RAxML and PhyML using similar computational runtime.
- Assesses branch support via UFBoot at speeds 10 to 40 times faster than RAxML rapid bootstrap and 100 times faster than non-parametric bootstrap, with reduced support bias.
- Includes ModelFinder, which selects substitution models and partitioning schemes 10 to 100 times faster than jModelTest and ProtTest.
- Provides built-in checkpointing to recover and resume analyses that have been prematurely halted.
- Supports broad model flexibility, including empirical and customizable mixture models, polymorphism-aware models, and partition models with mixed data types.
Watch out for
- Scaling to datasets with millions of sequences requires specialised workflows such as the CMAPLE module rather than standard maximum-likelihood tree search routines.
- Full-featured use of advanced models across massive alignments requires multicore or distributed high-performance computing resources.
Questions
Can I run IQ-TREE online without installing it?
IQ-TREE is preinstalled on the machine behind every research conversation. Ask the agent in plain language, or give it the exact command — it runs it, reads the output and explains it.
What licence is IQ-TREE under?
IQ-TREE is published under its authors' own licence; the project page at https://iqtree.github.io/ states it.
What alignment file formats does IQ-TREE accept?
IQ-TREE accepts standard multiple sequence alignment formats including FASTA, PHYLIP, Nexus, Clustal, and MSF.
How are tree outputs viewed or parsed?
Trees are exported in standard NEWICK format (.treefile), which can be visualised in viewer software such as FigTree, Dendroscope, or iTOL. A human-readable summary of the run and model parameters is also generated (.iqtree).
Can interrupted IQ-TREE analyses be resumed?
Yes. IQ-TREE automatically performs checkpointing during execution, enabling interrupted analyses to be continued from the saved state.
What options exist for assessing branch confidence?
Researchers can compute ultrafast bootstrap values (UFBoot), single-branch support metrics such as the SH-aLRT and aBayes tests, or tree topology tests including the approximately unbiased (AU) test.
How does IQ-TREE address site rate variation and ascertainment bias?
It accommodates rate heterogeneity across sites for all common data types and includes ascertainment bias correction for discrete character sets such as SNP data.
How to cite it
T.K.F. Wong, N. Ly-Trong, H. Ren, P. Demotte, H. Banos, A.J. Roger, E. Susko, C. Bielow, N. De Maio, N. Goldman, M.W. Hahn, M. dos Reis, L.S. Vinh, G. Huttley, R. Lanfear, B.Q. Minh (2026) IQ-TREE 3: Phylogenomic Inference Software using Complex Evolutionary Models. Mol. Biol. Evol., msag117.
Written from IQ-TREE's own README and project page. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.