Conda and Bioconda Tutorial: Install Bioinformatics Tools in Environments

Chapter 3 of 5: Creating and using environments

Now the useful part. In this chapter you will build an environment called qc with FastQC and MultiQC (the tools of the next tutorial in this path), use it, change it and delete it safely.

Step 5: Create your first environment

conda create -n qc fastqc multiqc

• conda create: make a new, empty environment and install packages into it.
• -n qc: the environment name (--name). Use short, meaningful names: qc, align, rnaseq, proj-wheat-2026.
• fastqc multiqc: the packages to install. You can list many in one command, and you should: one solve is faster and more reliable than many small installs.

conda first shows its plan and asks before changing anything:

Read the plan before typing y. Each line shows channel/subdir::name-version-build. Check that the bioinformatics tools come from bioconda and their dependencies from conda-forge. Notice that conda added Java (openjdk) for FastQC and Python for MultiQC automatically. Press y and Enter. Downloading and installing takes about 1–5 minutes, depending on your internet.

Add -y to skip the question in scripts: conda create -n qc -y fastqc multiqc.

Step 6: Activate the environment and run a tool

conda activate qc
which fastqc
fastqc --version
multiqc --version

Three signs that it worked: the prompt now starts with (qc); which fastqc points inside envs/qc/bin/; and the version commands answer. When you are done, leave the environment:

conda deactivate

Remember: every new terminal starts outside your environment. Activate it again each time with conda activate qc. Forgetting this is the number one cause of “command not found” in the next tutorials.

# conda activate qc (simplified)
if another environment is active:
    run its etc/conda/deactivate.d/ scripts
    remove its bin/ folder from PATH
PATH = "<prefix>/bin:" + PATH          # prefix = ~/miniforge3/envs/qc
CONDA_PREFIX = "<prefix>"
CONDA_DEFAULT_ENV = "qc"
run every script in <prefix>/etc/conda/activate.d/
PS1 = "(qc) " + PS1                    # the prompt change you see

Step 7: List, inspect and change environments

These are the commands you will use every week.

conda env list                      # all environments; * marks the active one
conda list -n qc                    # every package in qc, with version, build and channel
conda list -n qc fastqc             # just one package
conda install -n qc samtools        # add a tool to qc
conda update -n qc fastqc           # update one tool (within what the solver allows)
conda remove -n qc samtools         # remove a tool

• -n qc targets an environment by name, so you do not need to activate it first. Without -n, commands act on the active environment.
• conda list output has four columns: Name, Version, Build, Channel. The Channel column is the quickest way to check that nothing came from defaults.
• conda update --all -n qc updates everything in the environment. Do this at the start of a project, never in the middle of an analysis.

To run one command in an environment without activating it (handy in scripts and job files):

conda run -n qc fastqc --version

Step 8: Find packages and pin versions

Before installing, check that a tool exists and which versions are available:

conda search samtools
conda search "samtools>=1.20"

The output lists every available version and build, with the channel. You can also search on the web at bioconda.github.io (each tool has a recipe page) or anaconda.org/bioconda. Note that the conda package name is not always the program name. For example, the blast package provides blastn, blastp and friends.

To fix a version, add it to the package name. This is called pinning:

Put quotes around specs that contain >, < or |, otherwise the shell treats them as redirection or pipes.

Step 9: Remove, rename or clone an environment

conda deactivate                    # you cannot delete the environment you are in
conda env remove -n qc              # delete the environment and its folder
conda rename -n qc qc-old           # rename (recent conda versions)
conda create -n qc2 --clone qc      # copy an environment

Deleting an environment does not touch your data. It only removes the software folder. Never delete or move environment folders by hand with rm or mv; use these commands so conda’s records stay correct.

disk used ≈ size(package cache) + small per-environment overhead
not         number of environments × size of each environment

How many environments should I make?

The short answer: one environment per project or per pipeline step, never everything in base. Use this table as a guide.