Installing anvi'o on Mac OSX

The latest version of anvi’o is v8. See the release notes.

This page describes the anvi’o installation process for the current stable release on Mac OSX.

(1) Things you need before you start

You will need to run the installation commands from a terminal. Mac OSX comes with a basic Terminal application, or you can download and use a fancier one (such as iTerm).

Some of the packages we use need compiling during the installation process, so you should also make sure that you have Xcode Command Line Tools installed and up-to-date. Here is a quick link to their installation instructions. If you have to re-install the Command Line Tools, please remember to close your terminal window and open a new one before continuing with the anvi’o installation (a big thank you to Hilary Morrison for that tip).

You also need miniconda to be installed on your system. If you don’t already have it, please follow their installation instructions.

(2) Set up conda

Conda gives us a very simple and effective way to install anvi’o on your system along with most of its dependencies. To check if it was installed properly in the previous section, type conda in your terminal. You should see an output like this instead of a ‘command not found’ error (your version might be different):

$ conda --version
conda 23.7.4

Once you have confirmed you have conda installed, run this command to make sure you are up-to-date:

conda update conda

Good? Good! You are almost there!

(3) Setup an anvi’o environment

Working with Apple silicon

If you are using a computer with Apple silicon (like a M1 or M2 MacBook), you will find that some conda packages are not available (like bioconda packages). To avoid this issue, you can run the following command (only once) before creating the environment:

conda config --env --set subdir osx-64

It is a good idea to make sure you are not already in a conda environment before you run the following steps. Just to be clear, you can indeed install anvi’o in an existing conda environment, but if things go wrong, we kindly ask you to refer to meditation for help, rather than anvi’o community resources If you want to see what environments do you have on your computer and whether you already are in one of them in your current terminal by running conda env list. If all these are too much for you and all you want to do is to move on with the installation, simply do this: open a new terminal, and run conda deactivate, and continue with the rest of the text.

First, a new conda environment:

conda create -y --name anvio-8 python=3.10

And activate it:

conda activate anvio-8

Now you are in a pristine environment, in which you will install all conda packages that anvi’o will need to work properly. This looks scary, but it will work if you just copy-paste it and press ENTER:

conda install -y -c conda-forge -c bioconda python=3.10 \
        sqlite prodigal idba mcl muscle=3.8.1551 famsa hmmer diamond \
        blast megahit spades bowtie2 bwa graphviz "samtools>=1.9" \
        trimal iqtree trnascan-se fasttree vmatch r-base r-tidyverse \
        r-optparse r-stringi r-magrittr bioconductor-qvalue meme ghostscript \
        nodejs

# try this, if it doesn't install, don't worry (it is sad, but OK):
conda install -y -c bioconda fastani

If you see any error messages in the output indicating that a package failed to install, you should check the ‘Common problems’ section below or search for it in the anvi’o issues page (make sure to check the ‘Closed’ issues as well) to see if we already found a solution for the error.

(4) Install anvi’o

Here you will first download the Python source package for the official anvi’o release:

curl -L https://github.com/merenlab/anvio/releases/download/v8/anvio-8.tar.gz \
        --output anvio-8.tar.gz

Some packages in requirement.txt may require to be installed with a more up to date c-compiler on Mac OSX. Hence, we suggest all Mac users to run the following commands before you start the pip install command:

export CC=/usr/bin/clang
export CXX=/usr/bin/clang++

And then install it using pip like a boss:

pip install anvio-8.tar.gz

If you don’t see any error messages, then you are probably golden and can move on to testing your anvi’o setup in the section “Check your installation” :)

If you do see error messages, please know that you are not alone. We are as frustrated as you are. Please take a look at the problems people have reported and try these solutions, which will most likely address your issues. Common issues can be found on this page in the next section.

(5) Common problems

In this section, you will find the solutions to several common installation issues that users have reported. They are in no particular order, so it may help to search for the key words in your error messsage(s).

Can’t find your issue? Have you checked the anvi’o issues page, even the ‘Closed’ issues? If you can’t find a solution to your problem, please feel free to either open a new issue on our Github or send us a message on .

Issues with pysam installation using pip

Some people have reported errors in the installation of pysam using pip, so if your installation also fails due to pysam, you can use the following two lines to first install this package via conda, and then install the anvi’o package via pip:

conda install -y -c bioconda pysam
pip install anvio-8.tar.gz

At this point, you should probably test your samtools installation by running samtools --version. If you see an error that looks similar to this:

dyld: Library not loaded: @rpath/libcrypto.1.0.0.dylib
  Referenced from: /Users/iva/opt/miniconda3/envs/anvio-8/bin/samtools
  Reason: image not found
Abort trap: 6

This is happening because somehow you have the wrong version of the samtools :( The following commands should fix it:

conda remove -y samtools
conda install -y -c bioconda samtools=1.9

Then try samtools --version again to make sure it is okay now. What you should see is the following:

samtools 1.9
Using htslib 1.9
Copyright (C) 2018 Genome Research Ltd.

In anvi’o v8 there is a problem with incompatibility between the default installed versions of PyANI (v0.2.12) and matplotlib (v3.9.x). The program anvi-compute-genome-similarity can fail at the PyANI step, and the resulting log file will show an error like this:

AttributeError: module 'matplotlib.pyplot' has no attribute 'register_cmap'

To fix this issue, you can manually downgrade the matplotlib package like this:

pip install matplotlib==3.5.1

Or you can just switch to using FastANI instead. See this Discord thread for more context and details.

Issues with the C compiler

We realized that on some Mac OSX systems, some packages installed by pip requires a more up-to-date C compiler. If you’re getting an error that contains x86_64-apple-darwin13.4.0-clang or similar keywords in the output message, please run the following (which will set an environmental variable, and then try to install anvi’o via pip again):

export CC=clang
pip install anvio-8.tar.gz

If this didn’t work, try this more extensive solution:

export CC=/usr/bin/clang
export CXX=/usr/bin/clang++
pip install anvio-8.tar.gz

If the pip installation still doesn’t work (and especially if you see something like “clang-12: error: linker command failed with exit code 1” in the error message (we have often seen this error associated with the Levenshtein package), then this may be related to Xcode on Mac OSX. In this case you can try updating your Xcode by following the instructions described in this issue (in the “Solved it” section), and then try the pip command one more time.

If you did all that and it is still not working, please make an issue on the github page or let us know in the anvi’o Discord channel about your problem and we will try to help you.

(6) Check your installation

If you are here, you are ready to check if everything is working on your system. This section will help you finalize your installation so you are prepared for anything.

The easiest way to check your installation is to run the anvi’o program anvi-self-test:

anvi-self-test --suite mini

If you don’t want anvi’o to show you a browser window at the end and quietly finish testing if everything is OK, add --no-interactive flag to the command above. Another note, anvi-self-test is run in --suite mini mode, which tests the absolute minimal features of your anvi’o installation. If you run it without any parameters, it will tests many more things.

If everything goes smoothly, your browser should pop-up and show you an anvi’o interactive interface that looks something like this once anvi-self-test is done running:

The screenshot above is from 2015 and will be vastly different from the interactive interface you should see in your browser. It is still here so we remember where we came from 😇

If you are seeing the interactive interface, it means you now have a computer that can run anvi’o! In theory you can leave this page at this moment, but there are a few more details that would be best to attend now. So please bear with this tutorial just a little longer.

Don’t forget to come say hi to us on anvi’o Discord.

(6.1) Setup key resources

This is to further prepare your anvi’o installation for things you may need later, such as databases for taxonomic annotation of your genomes or functional annotation of your genes. This is an up-to-date list of programs that you should run in your terminal to have everything ready:

(6.2) Install an automated binning algorithm in your anvi’o environment

You can skip this section if you are not interested in reconstructing genomes from metagenomes using anvi’o.

Anvi’o offers a powerful interactive environment to reconstruct genomes from metageomes where you have full control over subtle decisions. For small assemblies (i.e., where you have less than 25,000 contigs), you do not need an additional binning software to reconstruct genomes from metagenomes. But for larger metagenomes, you have two options:

  • Use the program anvi-cluster-contigs with an automatic binning software that is already installed on your system.
  • Perform automatic binning outside of anvi’o, and import the binning results as a collection into anvi’o using the program anvi-import-collection to further refine those results.

The following recipe will help you install CONCOCT on your system just so there is an automatic binning algorithm ready on your system that you can use with anvi-cluster-contigs:

# setup a place to download CONCOCT source code
mkdir -p ~/github/ && cd ~/github/

# get a clone of the CONCOCT codebase from the fork
# that is tailored for the anvi'o conda environment
git clone https://github.com/merenlab/CONCOCT.git

# build and install
cd CONCOCT
python setup.py build
python setup.py install

Please note that you may encounter an error when running CONCOCT due to a TypeError. Please see the report #2154 for more information regarding this issue. IF you run into this issue, you may be able to resolve it by running the following command in your anvi’o conda environment: pip install scikit-learn==1.1.0. Georges Kanaan developed and tested this solution, and confirmed that it works at least for v8. But please let us know if this fix breaks any other part of anvi’o :)

If everything worked, when you type the following command,

anvi-cluster-contigs -h

You should see this output (where CONCOCT is found):

If you are a developer of an automatic binning algorithm and would like to see it in anvi’o, please get in touch with us. Anvi’o can pass any information about sequences (their coverages across samples, tetranucleotide frequencies, genes, functions, and whatever else you would like to have about them) to any program to run it on user data and import the results into anvi’o databases seamlessly through simple Python wrappers. Here are some examples of such wrappers for CONCOCT, for BinSanity, and for MaxBin2. If you wish to create one but are not sure how to test it, please start a GitHub issue.

(6.3) Troubleshooting

If your browser didn’t show up, or testing stopped with errors, please take a look at the common problems others have reported and try these solutions. Please remember you can always come to anvi’o Discord to ask for help if things are not working for you and the answers you find here are no use.

I see a lot of warning messages

It is absolutely normal to see ‘warning’ messages. In general anvi’o is talkative as it would like to keep you informed. In an ideal world you should keep a careful eye on those warning messages, but in most cases they will not require action.

If anvi-self-test fails with an error message that looks something like this,

libcrypto.so.1.0.0: cannot open shared object file: no such file or directory

it is likely that the pysam module installation failed. To fix this you should revisit the installation instructions, especially the part that says “Issues related to samtools”, and then come back to testing.

My browser didn’t show up

If your browser does not show up, or does show up but can’t show anything due to a ‘network problem’, you may also want to visit the address http://localhost:8080 by manually entering this address to your browser’s address bar, which should work on your local computer. On some systems the default network interface anvi’o uses to connect to its own server causes issues. You may also find the help page for anvi-interactive useful for future references.

If your browser does not show up while you are connected to a remote computer, it is quite normal. In some cases a text-based browser may show up instead of your graphical browser, too. This is becasue you are running anvi’o on another computer, and it tries to open a browser there. You can set things up for anvi’o to use your local browser to access to an anvi’o interactive interactive interface running remotely. For that, you can read this article (or ask your systems administrator to read it) to learn how you can forward displays from servers to your personal computer.

Browser shows up, but anvi’o complains about Chrome

If you are not using Chrome as your default browser, anvi’o will complain about it :/ We hate the idea of asking you to change your browser preferences for anvi’o :( But currently, Chrome maintains the most efficient SVG engine among all browsers we tested as of 2021. For instance, Safari can run the anvi’o interactive interface, however it takes orders of magnitude more time and memory compared to Chrome. Firefox, on the other hand, doesn’t even bother drawing anything at all. Long story short, the anvi’o interactive interface will not perform optimally with anything but Chrome. So you need Chrome. Moreover, if Chrome is not your default browser, every time interactive interface pops up, you will need to copy-paste the address bar into a Chrome window.

You can learn what is your default browser by running this command in your terminal:

python -c 'import webbrowser as w; w.open_new("http://")'

Everything is fine, but I can’t find anvi’o commands in a new terminal

If you open a new terminal and get command not found error when you run anvi’o commands, it means you need to activate anvi’o conda environment by running the following command (assuming that you named your conda environment for anvio as anvio-8, but you can always list your conda environments by running conda env list):

conda activate anvio-8

If you are getting an error that goes like,

Config Error: Something went wrong during the functional enrichment analysis :( We don't know
              what happened, but this log file could contain some clues: (...)

it often means that the R libraries that are needed to run functional enrichment analyses are not installed properly through conda :/ Luckily, you can try to install them using the R terminal as Marco Gabrielli shared on anvi’o Discord. For this, try running this command in your terminal:

Rscript -e 'install.packages(c("stringi", "tidyverse", "magrittr", "optparse"), repos="https://cloud.r-project.org")'

If everything goes alright, you can quit the R terminal by pressing CTRL+D twice. Once you are out, you can run this command to see if everything runs smoothly:

Rscript -e "library('tidyverse')"

In some cases the problem is the qvalue package, which can be a pain to install. If you are having hard time with that one, you can try this and see if that solves it:

Rscript -e 'install.packages("BiocManager", repos="https://cran.rstudio.com"); BiocManager::install("qvalue")'

Now you can take a look up some anvi’o resources here, or join anvi’o Discord to be a part of our growing community.

Other installation options

You will always find the official archives of anvi’o code as at the bottom of our GitHub releases as anvio-X.tar.gz:

https://github.com/merenlab/anvio/releases/latest

The best way to see what additional software you will need running on your computer for anvi’o to be happy is to take a look at the contents of this conda recipe (which is a conda build recipe, but it will give you the idea (ignore anvio-minimal, you basically have that one taken care of when you have anvi’o installed)).

Don’t be a stranger, and let us know if you need help through .

If you find a mistake on this page or would you like to update something in it, please feel free to edit its source by clicking the edit button at the top-right corner (which you will see if you are logged in to GitHub) 😇

Copyleft, The anvi'o Project