Iqtree Manual 1.0
User Manual: Pdf
Open the PDF directly: View PDF 
.
Page Count: 23
- Introduction
- Installation
- Tutorial
- First running example(Update!)
- Choosing the substitution model
- Support for phylogenetic likelihood library (New in version 1.0!)
- Novel tree search algorithm (New in version 1.0!)
- Codon models (New in version 1.0!)
- Morphological and SNP data (New in version 1.0!)
- Assessing branch supports with ultrafast bootstrap approximation
- Partitioned analysis for multi-gene alignments
- Utilizing multi-core CPUs
- Advanced tutorial
- Frequently asked questions (FAQ)
- Version History
IQ-TREE version 1.0 (July 2014)
Fast phylogenetic inference and ultrafast bootstrap
analysis by maximum likelihood.
User Manual and Tutorial
Please read carefully before using IQ-TREE the first time!
Project managers:
Bui Quang Minh - minh.bui(at)mfpl.ac.at
Arndt von Haeseler - arndt.von.haeseler(at)mfpl.ac.at
Core developers:
Lam-Tung Nguyen - tung.nguyen(at)mfpl.ac.at
Olga Chernomor - olga.chernomor(at)mfpl.ac.at
Diep Thi Hoang - diep.thi.hoang(at)gmail.com
Support:
Heiko A. Schmidt - heiko.schmidt(at)mfpl.ac.at
Contact address:
Center for Integrative Bioinformatics Vienna (CIBIV)
Max F. Perutz Laboratories, University of Vienna, Medical University of Vienna
Dr. Bohr-Gasse 9, A-1030 Vienna, Austria
License Agreement
This program is free software; you can redistribute it and/or modify it under the terms of the
GNU General Public License as published by the Free Software Foundation; either version
2 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WAR-
RANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
A PARTICULAR PURPOSE. See the GNU General Public License for more details.
Contents
1 Introduction 3
1.1 What’s new in version 1.0? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Key features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Installation 5
2.1 Binary release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Building source package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Building multi-core parallel version (Update!).................. 6
3 Tutorial 7
3.1 First running example(Update!).......................... 7
3.2 Choosing the substitution model . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.3 Support for phylogenetic likelihood library (New in version 1.0!) ....... 9
3.4 Novel tree search algorithm (New in version 1.0!)................ 9
3.5 Codon models (New in version 1.0!) ....................... 9
3.6 Morphological and SNP data (New in version 1.0!) ............... 10
3.7 Assessing branch supports with ultrafast bootstrap approximation . . . . . . 11
3.7.1 Assessing branch supports with standard nonparametric bootstrap . . 12
3.7.2 Assessing branch supports with single branch tests . . . . . . . . . . . 12
3.8 Partitioned analysis for multi-gene alignments . . . . . . . . . . . . . . . . . . 13
3.8.1 Choosing the right partitioning scheme . . . . . . . . . . . . . . . . . . 14
3.8.2 Bootstrapping with partition model . . . . . . . . . . . . . . . . . . . 15
3.9 Utilizing multi-core CPUs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4 Advanced tutorial 16
4.1 Tree topology tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.2 User-defined substitution models . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.3 Consensus construction and bootstrap value assignment . . . . . . . . . . . . 18
1
4.4 Computing Robinson-Foulds distance between trees . . . . . . . . . . . . . . . 19
4.5 Generating random trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5 Frequently asked questions (FAQ) 20
5.1 How does IQ-TREE treat gap/missing characters? . . . . . . . . . . . . . . . 20
6 Version History 20
2
1 Introduction
IQ-TREE is an efficient program for reconstructing large maximum likelihood trees and
assessing branch supports with the ultrafast bootstrap approximation. IQ-TREE extends
the IQPNNI algorithm with many enhancements. IQ-TREE is open-source and available
free of charge from
http://www.cibiv.at/software/iqtree/
IQ-TREE has been tested on Unix, Mac OS X, and Windows. The code of IQ-TREE has
been written in standard C/C++, which is possibly compilable on other platforms. Please
read the Installation section 2 for more details. We suggest that this documentation should
be read before using IQ-TREE the first time!
For impatient users we established a very user-friendly web server:
http://iqtree.cibiv.univie.ac.at
Its intuitive web interface allows users to perform online tree reconstruction within a few
clicks. Note that this online service only allows max. 12 CPU hours and 1 GB memory
per job. In case your job exceeds these limits, you can copy and paste the command-line
displayed to run the analysis at your local machine.
To cite IQ-TREE please use the following paper:
Bui Quang Minh, Minh Anh Thi Nguyen, and Arndt von Haeseler (2013) Ultrafast
approximation for phylogenetic bootstrap. Mol. Biol. Evol., 30:1188-1195.
Further readings on the methods developed:
•Heiko A. Schmidt and Arndt von Haeseler (2009) Phylogenetic Inference Using
Maximum Likelihood Methods. In P. Lemey, M. Salemi, A.M. Vandamme (eds.)The
Phylogenetic Handbook: a Practical Approach to Phylogenetic Analysis and Hypothesis
Testing., 2nd Edition, 181-209, Cambridge University Press, Cambridge.
•Bui Quang Minh, Le Sy Vinh, Arndt von Haeseler and Heiko A. Schmidt
(2005) pIQPNNI: Parallel reconstruction of large maximum likelihood phylogenies.
Bioinformatics, 21(19):3794-6.
•Le Sy Vinh and Arndt von Haeseler (2004) IQPNNI: Moving fast through tree
space and stopping in time. Mol. Biol. Evol., 21(8):1565-1571.
If you encounter bugs please send the .log file of the run and possibly the alignment to:
tung.nguyen(AT)univie.ac.at and minh.bui(AT)univie.ac.at.
3
1.1 What’s new in version 1.0?
Version 1.0 is the major release of the IQ-TREE software. We are happy to announce the
following new features:
•Integration of the phylogenetic likelihood library (PLL; Flouri et al., 2014) for fast
likelihood computation. This is enabled via -pll option and gives a speedup of 2X to
8X.
•A novel fast and effective stochastic algorithm for estimating maximum likelihood
phylogenies. It outperforms RAxML and PhyML in terms of log-likelihoods while
requiring similar amount of computing time. A manuscript describing the new method
was submitted. See Section 3.4 for more details.
•Codon models: GY (Goldman & Yang 1994), MG (Muse & Gaut 1994), and ECM
(Kosiol et al. 2007)
•Morphological models: MK and ORDERED (Lewis 2001)
•Ascertainment bias correction model (+ASC) for e.g., morphological or SNP data
(Lewis 2001)
•Nearest neighbor interchange with five (instead of one) branch optimization (-nni5)
is now the default option because of its higher accuracy
•SH-aLRT branch test also works now for partition models.
1.2 Key features
IQ-TREE provides a lot of options for phylogenetic reconstruction. The main features in-
clude:
•Reconstruction of the maximum likelihood tree from sequence alignments (Vinh and von Haeseler,
2004; Minh et al., 2005).
•Ultrafast bootstrap approximation for assessing branch supports (Minh et al., 2013).
•Various substitution models for binary, nucleotide, amino-acid with/without rate het-
erogeneity.
•Partition models for phylogenomic data
•Automatic selection of best-fit models similarly to ModelTest (Posada and Crandall,
1998).
•Standard non-parametric bootstrap (Felsenstein, 1985).
4
•Single branch tests (LBP, SH-aLRT; Adachi and Hasegawa, 1996; Guindon et al.,
2010).
•Test of model homogeneity assumption along the tree (Weiss and von Haeseler, 2003).
•Site-specific rate model (Meyer and von Haeseler, 2003).
•Fast consensus tree reconstruction, Robinson-Foulds distance computation.
2 Installation
See below for information how to install/build the different versions of the IQ-TREE software.
Executable versions of the sequential, that is, non-parallel program are intended for a number
of operating systems.
2.1 Binary release
1. Download the executable version of IQ-TREE for your operating system if it is avail-
able (iqtree-XXX-OS.tar.gz or iqtree-XXX-OS.zip, where XXX is the current version
number and OS the operating system) from
http://www.cibiv.at/software/iqtree
2. Extract the files (e.g., with tar xvzf iqtree-XXX-OS.tar.gz under Unix). This
should create a directory iqtree-XXX-OS.
3. You will find the executable in iqtree-XXX-OS/. This executable you should rename
to iqtree (or iqtree.exe on Windows systems) and copy it to your system search
path such that it is found by your system.
Note on multi–ompcore version: The executable is named iqtree-omp (or iqtree-omp.exe
on Windows). Please also copy other files needed for OpenMP (e.g., *.dll on Windows) to
the same folder that you copied iqtree-omp to. Finally, for Mac OS X you have to install
MacPorts and the associated gcc47 to run iqtree-omp properly (see a how-to in section 2.3).
If you encounter problems, please ask your local administrator for help.
2.2 Building source package
To build IQ-TREE from the sources you need a C++ compiler (e.g., gcc) and the CMake
tool installed (This is usually the case on UNIX/Linux systems. For Windows you might
want to obtain CygWin/MinWG/MS Visual C++ or XCode for MacOSX). Then you can
follow the procedure below:
5
1. Download the current version of the software (iqtree-XXX-Source.tar.gz or
iqtree-XXX-Source.zip, where XXX is the current version number) from
http://www.cibiv.at/software/iqtree
2. Extract the files (e.g., with tar xvzf iqtree-XXX-Source.tar.gz under Unix). This
should create a directory iqtree-XXX-Source.
3. Change into this directory.
4. Create a sub-directory build and go into this sub-directory by entering:
mkdir build
cd build
5. Configure the source codes using CMake:
cmake ..
6. Compile and build the source codes:
make
This creates an executable iqtree (or iqtree.exe on Windows systems). This exe-
cutable can copied to your system search path such that it is found by your system.
If you encounter problems, please ask your local administrator for help.
2.3 Building multi-core parallel version (Update!)
To build the multi-core version you need a compiler that supports the OpenMP standard
(e.g., gcc). For Linux and Windows the gcc and MinGW compilers work just fine. However,
in our test on Mac OS X, IQ-TREE was successfully compiled with the default the XCode gcc
but the example run crashed for unknown reason. Therefore, we employed MacPorts (with
gcc47 or later) and successfully ran IQ-TREE compiled with MacPorts gcc. To this end,
please first install MacPorts, gcc in MacPorts (sudo port install gcc47) and configure
gcc to point to the MacPorts’ gcc version (sudo port select --set gcc mp-gcc47).
The compilation then follows the same route with slightly changed command line for cmake:
cmake .. -DIQTREE_FLAGS="omp"
All other commands remain the same. It is recommended to copy the executable file
iqtree-omp (or iqtree-omp.exe on Windows) to the system search path such that one
can simply run iqtree-omp from the command-line.
6
3 Tutorial
This section gives users a quick starting guide. You can either download the binary for
your platform from the IQ-TREE website or the source code. In the later case, you will
need to compile the source code (see Installation section 2). For the next steps, the iqtree
executable should be then copied into the bin folder such that IQ-TREE can be invoked by
simply entering ’iqtree’ at the command-line. You can run ’iqtree -h’ to see a list of
options available in IQ-TREE.
3.1 First running example(Update!)
From the download there is an example alignment called example.phy in PHYLIP format
(IQ-TREE also supports FASTA and NEXUS files). You can now start to reconstruct a
maximum-likelihood tree from this alignment by typing (assuming that you are now in the
same folder with example.phy):
iqtree -s example.phy
’-s’ is the option to specify the name of the alignment file that is always required by
IQ-TREE to work. At the end of the run IQ-TREE will write several output files:
•example.phy.iqtree: the main report file that is self readable for users. You should
look at this file to see the results.
•example.phy.treefile: the ML tree in NEWICK format, which can be visualized by
tree viewer tools such as FigTree, iTOL. Note that this newick tree is also embedded
in example.phy.iqtree.
•example.phy.log: log file of the entire run (also printed on the screen). To report
bugs, please send this log file and the original alignment file to the authors.
Note that all output files have the default prefix as the alignment file name. You can always
change the prefix using the ’-pre’ option, e.g.:
iqtree -s example.phy -pre myprefix
Then IQ-TREE will write output files myprefix.iqtree, myprefix.treefile, etc. This
is helpful when you do several runs for the same input.
******** NEW IN VERSION 1.0 ********
Since version 1.0 IQ-TREE by default offers a more accurate tree search and bootstrap by
optimizing five branches around the nearest neighbor interchanges (NNIs). This comes with
a trade-off of approximately 2X longer running time than 0.9.X version. To switch back to
old behaviour of optimizing one branch around NNIs, simply use the -nni1 option:
7
iqtree -s example.phy -nni1
3.2 Choosing the substitution model
IQ-TREE supports numerous substitution models for binary, DNA, and protein data and
Gamma rate heterogeneity model. If you do not specify, IQ-TREE will use the default HKY,
WAG, and JC models for DNA, protein, and binary alignments, respectively. For most data
sets these models are too simplified. If you have no idea about which model is appropriate
for your data, let IQ-TREE automatically determine the best-fit model for your alignment
with:
iqtree -s example.phy -m TEST
’-m’ is the option to specify the model name to use during the analysis. ’TEST’ is a key
word telling IQ-TREE to first select the best-fit model. The remaining analysis will be done
using the selected model. More specifically, IQ-TREE computes the log-likelihoods of the
initial BIONJ tree for many different models and the Akaike information criterion (AIC),
corrected Akaike information criterion (AICc), and the Bayesian information criterion (BIC).
Then IQ-TREE chooses the model that minimizes the BIC score (you can also change to
AIC or AICc by adding the option ”-AIC” or ”-AICc”, respectively). Moreover, IQ-TREE
will write an additional file:
•example.phy.model: log-likelihoods for all models tested.
If you now look at example.phy.iqtree you will see that IQ-TREE selected the model
’TIM’ with ’Invar+Gamma’ rate heterogeneity. So ’TIM+I+G’ is the best-fit model for this
example data. From now on you can run with e.g.:
iqtree -s example.phy -m TIM+I+G
Sometimes you only want to find the best-fit model without doing tree reconstruction, then
run:
iqtree -s example.phy -m TESTONLY
Here, IQ-TREE will stop after finishing the model selection. The name of the best-fit model
will be printed on the screen. Finally, note that IQ-TREE will check if the file ’*.model’
exists and is correct. If so, it will automatically reuse the log-likelihoods computed to speed
up the model selection procedure.
********NEW********
Since version 0.9.6 IQ-TREE offers the partition model selection for multi-gene analysis. See
Section 3.8.1 for more details.
8
3.3 Support for phylogenetic likelihood library (New in version 1.0!)
In the major release 1.0, we added the support for PLL (Flouri et al., 2014) which helps
to speed up IQ-TREE by a factor of 2X to 8X. To test the new feature simply use option
’-pll’, for example:
iqtree -s example.phy -pll -m GTR+G
Here, we also specifies model GTR+G as it is currently supported by PLL. Note that this
option does not entirely work with other options yet (such as ’-m TEST’). In such cases, an
error message will be displayed.
3.4 Novel tree search algorithm (New in version 1.0!)
IQ-TREE 1.0 implemented a new tree search algorithm, which explore the tree space much
more efficiently than version 0.9.X. Here, IQ-TREE combines parsimony analysis to provide
better starting trees, new stochastic algorithm to escape local optima, and new stopping rule.
The new search strategies come with a few parameters where the default values were tested
to work well for many different data sets. Moreover, you can change the default parameters
with options:
-numpars <number> Number of initial parsimony trees (default: 100)
-toppars <number> Number of top initial parsimony trees (dfault: 20)
-numcand <number> Number of candidate trees during search (defaut: 5)
-pers <perturbation> Perturbation strength for stochastic NNI (default: 0.5)
-numstop <number> Number of unsuccessful iterations to stop (default: 100)
Finally, you can still switch back to the old algorithm of 0.9.X by options:
-iqp Use IQP tree perturbation (default: sNNI)
-iqpnni Switch entirely to old IQPNNI algorithm
3.5 Codon models (New in version 1.0!)
IQ-TREE 1.0 supports basic codon models (GY, MG, and ECM). You need to input a
protein-coding DNA alignment and specify codon data by option ’-st CODON’ (Otherwise,
IQ-TREE applies DNA model because it detects that your alignment has DNA sequences):
iqtree -s coding_gene.phy -st CODON
9
If your alignment length is not divisible by 3, an error message will occur. IQ-TREE will
group sites 1,2,3 into codon site 1; site 4,5,6 to codon site 2; etc. Moreover, any codon,
which has at least one gap/unknown/ambiguous nucleotide, will be treated unknown codon
character.
If you are not sure which model to use, simply add ’-m TEST’, which also works for codon
alignments:
iqtree -s coding_gene.phy -st CODON -m TEST
By default IQ-TREE uses the standard genetic code. You can change to other genetic
code (see http://www.ncbi.nlm.nih.gov/Taxonomy/Utils/wprintgc.cgi) with following
options:
Option Genetic code
-st CODON1 The Standard Code (same as -st CODON)
-st CODON2 The Vertebrate Mitochondrial Code
-st CODON3 The Yeast Mitochondrial Code
-st CODON4 The Mold, Protozoan, and Coelenterate Mitochondrial Code and
the Mycoplasma/Spiroplasma Code
-st CODON5 The Invertebrate Mitochondrial Code
-st CODON6 The Ciliate, Dasycladacean and Hexamita Nuclear Code
-st CODON9 The Echinoderm and Flatworm Mitochondrial Code
-st CODON10 The Euplotid Nuclear Code
-st CODON11 The Bacterial, Archaeal and Plant Plastid Code
-st CODON12 The Alternative Yeast Nuclear Code
-st CODON13 The Ascidian Mitochondrial Code
-st CODON14 The Alternative Flatworm Mitochondrial Code
-st CODON16 Chlorophycean Mitochondrial Code
-st CODON21 Trematode Mitochondrial Code
-st CODON22 Scenedesmus obliquus Mitochondrial Code
-st CODON23 Thraustochytrium Mitochondrial Code
-st CODON24 Pterobranchia Mitochondrial Code
-st CODON25 Candidate Division SR1 and Gracilibacteria Code
3.6 Morphological and SNP data (New in version 1.0!)
IQ-TREE 1.0 supports discrete morphological alignment by ’-st MORPH’ option:
iqtree -s morphology.phy -st MORPH
IQ-TREE implements to two morphological ML models (MK and ORDERED; see Lewis
2001), where MK is the default model. MK is a Juke-Cantor-like model. ORDERED model
10
considers only transitions between states i→i−1, i→i, and i→i+ 1. Morphological
data typically do not have constant (uninformative) sites. In such case, you should apply
ascertainment bias correction model by e.g.:
iqtree -s morphology.phy -st MORPH -m MK+ASC
You can again select best-fit model with ’-m TEST’ (which also consider +G):
iqtree -s morphology.phy -st MORPH -m TEST
For SNP data (DNA) that typically do not contain constant sites, you can explicitly tell
model to include ascertainment bias correction:
iqtree -s SNP_data.phy -m GTR+ASC
You can explicitly tell model testing to only include ’+ASC’ model with:
iqtree -s SNP_data.phy -m TEST+ASC
3.7 Assessing branch supports with ultrafast bootstrap approximation
The ultrafast bootstrap approximation is the most value-added feature available in IQ-
TREE. Simply run:
iqtree -s example.phy -m TIM+I+G -bb 1000
’-bb’ specifies the number of bootstrap replicates where 1000 is the minimal number recom-
mended. When you now look at the section ’MAXIMUM LIKELIHOOD TREE’ in example.phy.iqtree,
you will see that every internal node of the tree figure will be associated a support value in per-
centage. Branch supports are assigned onto the ML tree and printed in example.phy.treefile
that can be viewed again in FigTree. In addition, IQ-TREE writes the following files:
•example.phy.contree: the consensus tree with assigned branch supports where branch
lengths are optimized on the original alignment.
•example.phy.splits: support values in percentage for all splits (bipartitions), com-
puted as the occurence frequencies in the bootstrap trees. This file is in ”star-dot”
format.
•example.phy.splits.nex: has the same information as example.phy.splits but in
NEXUS format, which can be viewed with SplitsTree program.
11
3.7.1 Assessing branch supports with standard nonparametric bootstrap
The standard nonparametric bootstrap can be invoked by:
iqtree -s example.phy -m TIM+I+G -b 100
’-b’ specifies the number of bootstrap replicates where 100 is the minimal number recom-
mended. IQ-TREE will additionally writes the following files:
•example.phy.boottrees: the set of bootstrap trees reconstructed.
•example.phy.contree: the bootstrap consensus tree with assigned branch supports
where branch lengths are optimized on the original alignment.
3.7.2 Assessing branch supports with single branch tests
IQ-TREE provides an implementation of the SH-like approximate likelihood ratio test (SH-
aLRT; Guindon et al., 2010). To perform this test, simply run:
iqtree -s example.phy -m TIM+I+G -alrt 1000
’-alrt’ specifies the number of bootstrap replicates for SH-aLRT where 1000 is the minimal
number recommended. IQ-TREE will perform SH-aLRT at the end of the tree reconstruction
process and assign support values onto the ML tree. The support values will be reflected in
the tree file example.phy.treefile.
IQ-TREE also provides a fast implementation of the local bootstrap probabilities method
(Adachi and Hasegawa, 1996), which we call Fast-LBP. Fast-LBP computes the branch sup-
port by comparing the tree log-likelihood with the log-likelihoods of the two alternative
nearest-neighbor-interchange (NNI) trees around the branch of interest. However, Fast-LBP
is different from LBP where we compute the log-likelihoods of the two alternative NNI trees
by only reoptimizing five branches around the branch of interest (Similar idea is used in the
SH-aLRT test). To perform Fast-LBP, simply run:
iqtree -s example.phy -m TIM+I+G -lbp 1000
You can also perform both tests:
iqtree -s example.phy -m TIM+I+G -alrt 1000 -lbp 1000
The branches of the resulting ML tree will be assigned with both SH-aLRT and Fast-LBP
support values. Finally, you can also combine the ultrafast bootstrap approximation with
single branch tests within one single run:
iqtree -s example.phy -m TIM+I+G -bb 1000 -alrt 1000 -lbp 1000
12
3.8 Partitioned analysis for multi-gene alignments
In the partition model, you can specify a substitution model for each gene/character set
individually. IQ-TREE will then estimate the model parameters and branch lengths sepa-
rately for every partition. To this end, you have to first prepare a NEXUS file including a
SETS block with CharSet and CharPartition commands to specify individual genes and the
partition, respectively. For example:
#nexus
begin sets;
charset part1 = 1-100;
charset part2 = 101-384;
charpartition mine = HKY+G:part1, GTR+I+G:part2;
end;
Now if you save this into a file example.nex and run:
iqtree -s example.phy -sp example.nex
This means that IQ-TREE will partition the alignment example.phy into 2 subsets named
part1 and part2 containing sites (columns) 1-100 and 101-384, respectively. Moreover, IQ-
TREE applies the subtitution models HKY+G and GTR+I+G to part1 and part2, respectively.
After the run has finished, the example.nex.iqtree file will contain substitution model
parameters, trees with branch lengths for all subsets in the partition.
Moreover, the CharSet command allows to specify non-consecutive sites using comma-
separated list of ranges with e.g.:
charset part1 = 1-100 200-384;
That means, part1 contains sites 1-100 and 200-384 of the alignment. Another example is:
charset part1 = 1-100\3;
for extracting sites 1,4,7,...,100 from the alignment. This is useful for getting codon positions
from the protein-coding alignment.
Moreover, IQ-TREE allows a more advanced feature compared to other programs: IQ-TREE
allows different subsets coming from different alignments. For example:
#nexus
begin sets;
charset part1 = part1.phy: 1-100\3 201-300\3;
13
charset part2 = part2.phy: 101-300;
charpartition mine = HKY:part1, GTR+G:part2;
end;
Here, part1 and part2 are read from alignment files part1.phy and part2.phy, respectively
(a ’:’ is needed to separate the alignment file name and site specification). Because the
alignment file names were embedded in this NEXUS file, you can simply run:
iqtree -sp example.nex
Note that part1.phy and part2.phy need not contain the same set of sequence names.
That means, if some sequence occurs in part1.phy but not in part2.phy, IQ-TREE will
treat corresponding part of sequence in part2.phy as missing data. For your convenience
IQ-TREE writes the concatenated alignment into the file example.nex.conaln.
********NEW EXPERIMENTAL FEATURE********
Since version 0.9.6 IQ-TREE supports partition models with joint and proportional branch
lengths between genes. This is to reduce the number of parameters in case of model overfitting
for the full partition model. For example:
iqtree -spp example.nex
applies a proportional partition model. That means, we have only one set of branch lengths
for species tree but allow each gene to evolve under a specific rate (scaling factor) normalized
to the average of 1.
A partition model with joint branch lengths is specified by:
iqtree -spj example.nex
(i.e., all gene-specific rates are equal to 1).
3.8.1 Choosing the right partitioning scheme
Since version 0.9.6 IQ-TREE implements a greedy strategy (Lanfear et al., 2012) that starts
with the full partition model and sequentially merges two genes until the model fit does not
increase any further:
iqtree -sp example.nex -m TESTLINK
After the best partition is found IQ-TREE will immediately start the tree reconstruction
under the best-fit partition model. Sometimes you only want to find the best-fit partition
model without doing tree reconstruction, then run:
14
iqtree -sp example.nex -m TESTONLYLINK
3.8.2 Bootstrapping with partition model
IQ-TREE can perform the ultrafast bootstrap with partition models by e.g.,
iqtree -sp example.nex -bb 1000
Here, IQ-TREE will resample the sites within subsets of the partitions (i.e., the bootstrap
replicates are generated per subset separately and then concatenated together). The same
holds true if you do the standard nonparametric bootstrap.
********NEW********
Since version 0.9.6 IQ-TREE supports the gene-resampling strategy:
iqtree -sp example.nex -bb 1000 -bspec GENE
is to resample genes instead of sites. Moreover, IQ-TREE allows an even more complicated
strategy: resampling genes and sites within resampled genes:
iqtree -sp example.nex -bb 1000 -bspec GENESITE
3.9 Utilizing multi-core CPUs
A specialized version of IQ-TREE allows users to perform the analysis that utilizes multiple
cores during the run (made possible by the OpenMP library). You can download the binary
from the software website or compile the source code yourself (see Installation section 2).
For the following please copy the binary iqtree-omp and other files in the package bin folder
into the system bin folder such that it can be invoked from the command-line by simply
running the command iqtree-omp.
If you now run with e.g.:
iqtree-omp -s example.phy
Then IQ-TREE will use all the available cores of your CPU. This might not be a good
practice because our parallelization technique only works well on long alignments. If you
have a very short alignment, it is not recommended to use this IQ-TREE version. Because
the speedup gain depends on the alignment length, a good practice is to run this version
with increasing number of cores by e.g.:
15
iqtree-omp -s example.phy -omp 2
Here, -omp is the option to specify the number of cores that IQ-TREE will use. If you see
that the wall-clock time reduction is substantial compared with the sequential IQ-TREE
version, then you can try:
iqtree-omp -s example.phy -omp 3
and so on, until no substantial reduction of running time is observed. The remaining analysis
can then be carried out with that number of cores.
For example, on my computer (Linux, Intel Core i5-2500K, 3.3 GHz, quad cores) I observed
the following wall-clock running time for this example alignment:
No. cores Wall-clock time
1 21.465 sec.
2 13.627 sec.
3 11.119 sec.
4 10.807 sec.
Therefore, I would only use 2 cores for this specific alignment ("-omp 2" option).
4 Advanced tutorial
This section gives an advanced tutorial for more experienced users. It includes several
advanced features like tree topology test, user-defined substitution models.
4.1 Tree topology tests
IQ-TREE can compute log-likelihoods of user-defined trees passed via -z option:
iqtree -s example.phy -z example.treels
assuming that example.treels contains the trees in NEWICK format. At the end of the
usual run, IQ-TREE will additionally evaluate all trees in there using the estimated model
parameters. When you look into example.phy.iqtree there will be a section USER TREES
that lists the tree IDs and the corresponding log-likelihoods. Moreover, IQ-TREE will addi-
tionally write a file:
•example.phy.treels.trees: the trees with optimized branch lengths.
16
If you only want to evaluate the trees without reconstructing the ML tree, you can run:
iqtree -s example.phy -z example.treels -n 1
Here, IQ-TREE will only reconstruct the BIONJ+NNI tree and use that tree to estimate
the model parameters, which are normally accurate enough for our purpose.
IQ-TREE also supports several tree topology tests using the RELL approximation (Kishino et al.,
1990) including: bootstrap proportion (BP), Kishino-Hasegawa test (KH; Kishino and Hasegawa,
1989), Shimodaira-Hasegawa test (SH; Shimodaira and Hasegawa, 1999), expected likeli-
hood weights (ELW; Strimmer and Rambaut, 2002), weighted-KH (WKH), and weighted-
SH (WSH) tests. The trees are passed via -z option, thus you can run:
iqtree -s example.phy -z example.treels -n 1 -zb 1000
Here, -zb specifies the number of RELL replicates, where 1000 is the minimum number
recommended. The USER TREES section of example.phy.iqtree will list the results of BP,
KH, SH, and ELW methods. If you want to also perform the WKH and WSH, simply add
-zw option:
iqtree -s example.phy -z example.treels -n 1 -zb 1000 -zw
Finally, note that IQ-TREE will automatically detect duplicated tree topologies and omit
them during the evaluation.
4.2 User-defined substitution models
Users can specify an arbitrary DNA models using a 6-letter specification that constrains
which rates to be equal. For example, 010010 corresponds to the HKY model and 012345
the GTR model. In fact, the IQ-TREE source code internally uses this specification to
simplify the coding. The 6-letter code is specified via -m option, e.g.:
iqtree -s example.phy -m 010010+G
Moreover, with -m option one can input a file name which contains the 6 rates (A-C, A-G,
A-T, C-G, C-T, G-T) and 4 base frequencies (A, C, G, T), e.g.:
iqtree -s example.phy -m mymodel+G
where mymodel is a file containing the 10 entries described above. One can even specify the
rates within -m option by e.g.:
17
iqtree -s example.phy -m ’TN{2.0,3.0}+G8{0.5}+I{0.15}’
That means, we use Tamura-Nei model with fixed transition-transversion rate ratio of 2.0 and
purine/pyrimidine rate ratio of 3.0. Moreover, we use an 8-category Gamma-distributed site
rates with the shape parameter (alpha) of 0.5 and a proportion of invariable sites p-inv=0.15.
Note that by default IQ-TREE computes empirical state frequencies from the alignment, but
one can also optimize the frequencies by maximum-likelihood with +Fo in the model name:
iqtree -s example.phy -m GTR+G+Fo
For amino-acid alignments, if one wants to use the frequencies of the empirical protein model,
then use +Fu, for example:
iqtree -s myprotein_alignment -m WAG+G+Fu
Finally, note that all model specifications above can be used in the partition model NEXUS
file.
4.3 Consensus construction and bootstrap value assignment
IQ-TREE can construct an extended majority-rule consensus tree from a set of trees written
in NEWICK or NEXUS format (e.g., produced by MrBayes):
iqtree -con mytrees
To build a majority-rule consensus tree, simply set the minimum support threshold to 0.5:
iqtree -con mytrees -t 0.5
If you want to specify a burn-in (the number of beginning trees to ignore from the trees file),
use -bi option:
iqtree -con mytrees -t 0.5 -bi 100
to skip the first 100 trees in the file.
IQ-TREE can also compute a consensus network and print it into a NEXUS file by:
iqtree -net mytrees
18
Finally, an useful feature is to read in an input tree and a set of trees, then IQ-TREE can
assign the support value onto the input tree (number of times each branch in the input tree
occurs in the set of trees) by:
iqtree -sup input_tree set_of_trees
4.4 Computing Robinson-Foulds distance between trees
IQ-TREE implements a very fast Robinson-Foulds (RF) distance computation using hash
table, which is a lot faster than PHYLIP package. For example, you can run:
iqtree -rf tree_set1 tree_set2
to compute the pairwise RF distances between 2 sets of trees. If you want to compute the
all-to-all RF distances of a set of trees, use:
iqtree -rf_all tree_set
4.5 Generating random trees
IQ-TREE provides several random tree generation models. For example,
iqtree -r 100 100.tree
is to generate a 100-taxon random tree into the file 100.tree under the Yule Harding model,
where the branch lengths follow an exponential distribution with mean of 0.1. If you want
to change the branch length distribution, run e.g:
iqtree -r 100 -rlen 0.05 0.2 0.3 100.tree
to set the minimum, mean, and maximum branch lengths as 0.05, 0.2, and 0.3, respectively.
If you want to generate trees under uniform model instead, use ’-ru’ option:
iqtree -ru 100 100.tree
19
5 Frequently asked questions (FAQ)
5.1 How does IQ-TREE treat gap/missing characters?
Gaps (-) and missing characters (? or N for DNA alignments) are treated in the same
way as unknown characters, which represent no information. The same treatment holds for
many other ML software (RAxML, PhyML, etc.). Technically in the Felsenstein’s pruning
algorithm we fill a partial likelihood vector of all 1’s for all character states. This is the same
as follows. For a site (column) of an alignment containing AC-AG-A (i.e. A for sequence 1, C
for sequence 2, - for sequence 3,...), the site-likelihood of a tree T is equal to the site-likelihood
of the subtree of T restricted to those sequences containing non-gap characters:
ℓ(T|AC −AG −A) = ℓ(Tsub|ACAGA)
6 Version History
Version 0.9.6: October 2013
•Ultrafast model selection and partitioning for phylogenomic alignments.
•Introduction of nearest neighbor interchange (NNI) with five branch optimization
to evaluate candidate NNIs. This will bring higher accuracy for tree reconstruction
and bootstrap with a tradeoff of c.a. 2X longer running time.
•Introduction of joint and proportional partition models to reduce the number of
parameters in case of model overfitting (experimental).
•Introduction of gene-resampling and gene-and-site resampling for the bootstrap
on multi-gene alignments.
Version 0.9.5: May 2013
•Introduction of bootstrap epsilon to select equally good bootstrap trees at random
to deal with polytomies
Version 0.9.4: Easter 2013
•Tree topology tests
Version 0.9.3: March 2013
•New implementation of model selection that works on all data types.
•A tutorial about using partition models.
•Parallel OpenMP support to utilize multi-core CPUs.
Version 0.9.0: September 2012 - First beta release.
20
Credits and Acknowledgement
Some parts of the code were taken from the following packages/libraries: Phylogenetic likeli-
hood library (Flouri et al., 2014), TREE-PUZZLE (Schmidt et al., 2002), BIONJ (Gascuel,
1997), Nexus Class Libary (Lewis, 2003), Eigen library (Guennebaud et al., 2010), SPRNG
library (Mascagni and Srinivasan, 2000), Zlib library (http://www.zlib.net).
Financial supports from the Austrian Science Fund (FWF), the Vienna Science and Tech-
nology Fund (WWTF), and the University of Vienna are greatly appreciated.
References
Adachi, J. and Hasegawa, M. (1996) MOLPHY Version 2.3 – Programs for Molecular Phy-
logenetics Based on Maximum Likelihood, volume 28 of Computer Science Monographs.
Institute of Statistical Mathematics, Minato-ku, Tokyo.
Felsenstein, J. (1985) Confidence limits on phylogenies: An approach using the bootstrap.
Evolution,39, 783–791.
Flouri, T., Izquierdo-Carrasco, F., Darriba, D., Aberer, A. J., Nguyen, L.-T., Minh, B. Q.,
von Haeseler, A. and Stamatakis, A. (2014) The Phylogenetic Likelihood Library. System-
atic Biology,63, accepted.
Gascuel, O. (1997) BIONJ: An improved version of the NJ algorithm based on a simple
model of sequence data. Mol. Biol. Evol.,14, 685–695.
Guennebaud, G., Jacob, B. et al. (2010) Eigen v3. http://eigen.tuxfamily.org.
Guindon, S., Dufayard, J.-F., Lefort, V., Anisimova, M., Hordijk, W. and Gascuel, O. (2010)
New algorithms and methods to estimate maximum-likelihood phylogenies: Assessing the
performance of phyml 3.0. Syst. Biol.,59, 307–321.
Kishino, H. and Hasegawa, M. (1989) Evaluation of the maximum likelihood estimate of
the evolutionary tree topologies from DNA sequence data, and the branching order in
Hominoidea. J. Mol. Evol.,29, 170–179.
Kishino, H., Miyata, T. and Hasegawa, M. (1990) Maximum likelihood inference of protein
phylogeny and the origin of chloroplasts. J. Mol. Evol.,31, 151–160.
Lanfear, R., Calcott, B., Ho, S. Y. W. and Guindon, S. (2012) PartitionFinder: Combined
selection of partitioning schemes and substitution models for phylogenetic analyses. Mol.
Biol. Evol.,29, 1695–1701.
Lewis, P. O. (2003) NCL: a C++ class library for interpreting data files in NEXUS format.
Bioinformatics,19, 2330–2331.
21
Mascagni, M. and Srinivasan, A. (2000) Algorithm 806: SPRNG: A scalable library for
pseudorandom number generation. ACM Transactions on Mathematical Software,26, 436–
461.
Meyer, S. and von Haeseler, A. (2003) Identifying site-specific substitution rates. Mol. Biol.
Evol.,20, 182–189.
Minh, B. Q., Nguyen, M. A. T. and von Haeseler, A. (2013) Ultrafast approximation for
phylogenetic bootstrap. Mol. Biol. Evol.,in press.
Minh, B. Q., Vinh, L. S., von Haeseler, A. and Schmidt, H. A. (2005) pIQPNNI – parallel
reconstruction of large maximum likelihood phylogenies. Bioinformatics,21, 3794–3796.
Posada, D. and Crandall, K. A. (1998) MODELTEST: Testing the model of DNA substitu-
tion. Bioinformatics,14, 817–818.
Schmidt, H. A., Strimmer, K., Vingron, M. and von Haeseler, A. (2002) TREE-PUZZLE:
Maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioin-
formatics,18, 502–504.
Shimodaira, H. and Hasegawa, M. (1999) Multiple comparisons of log-likelihoods with ap-
plications to phylogenetic inference. Mol. Biol. Evol.,16, 1114–1116.
Strimmer, K. and Rambaut, A. (2002) Inferring confidence sets of possibly misspecified gene
trees. Proc. R. Soc. Lond. B,269, 137–142.
Vinh, L. S. and von Haeseler, A. (2004) IQPNNI: Moving fast through tree space and stop-
ping in time. Mol. Biol. Evol.,21, 1565–1571.
Weiss, G. and von Haeseler, A. (2003) Testing substitution models within a phylogenetic
tree. Mol. Biol. Evol.,20, 572–578.
22
