Getting started


seq2science requires anaconda

Download and install miniconda if not yet installed:

user@comp:~$ wget -O
user@comp:~$ bash # (make sure to say **yes** when asked to initialize conda)
user@comp:~$ source ~/.bashrc

Set the correct channels (in this specific order) to use bioconda:

user@comp:~$ conda config --add channels defaults
user@comp:~$ conda config --add channels bioconda
user@comp:~$ conda config --add channels conda-forge

Running a workflow

A typical setup and run of a workflow looks like this, where you start with activating the seq2science environment.

(base) user@comp:~$ conda activate seq2science

Then navigate to your project dir.

(seq2science) user@comp:~$ cd my_project

Where you initialize the workflow with a configuration file and samples file, and edit those to your needs.

(seq2science) user@comp:~/my_project$ seq2science init {workflow}

And finally run the workflow. Note that the first time you run a workflow it will take a while before the real run starts, as seq2science will first install all the necessary software for the analysis steps.

(seq2science) user@comp:~.my_project$ seq2science run {workflow} --cores 24

Getting an explanation of what seq2science did (or will do)

Seq2science has a function to write an explanation of what has/will be done with the configuration file for a workflow:

(seq2science) user@comp:~.my_project$ seq2science explain {workflow}

This will print an extensive (quite technical) explanation that can also serve as a starting point for a material and methods section. This is not a replacement for the documentation, which we definitely recommend you to always read! This explanation is also automatically added to the final QC report.

Where does seq2science store results and looks for ‘starting points’?

We recommend that for a typical run of seq2science you use a folder structure like this:

└── my_project
    ├── samples.tsv
    ├── config.yaml
    └── {result_dir}
        └── {fastq_dir}
            ├── sample1.fastq.gz
            ├── sample2_R1.fastq.gz
            └── sample2_R2.fastq.gz

This structure is the default setting of seq2science, however can be adjusted to your liking in the config.yaml.

What is Snakemake?

Since under-the-hood seq2science is based on snakemake, we thought it might be nice to give a little introduction to snakemake.

Snakemake is a pipeline tool that allows users to specify rules. Each rule is defined with what it requires as input, what it will output, and what command it needs to run to generate the output from the input. This design allows for the linking of many rules, where the input of one rule is the output of another. When invoking Snakemake it will then decide itself which rules need to be run for the output you specified.

Here is an example Snakefile with just two (very simple) rules:

rule one:
        # for this example we just make an empty file
        "touch {output}"

rule two:
        # for this example we just make an empty file
        "touch {output}"

We can tell snakemake to generate for instance file1.txt like this:

(base) user@comp:~$ snakemake file1.txt

And snakemake will see that rule one can generate this output, sees that it requires no input, and executes the shell command. If we tell snakemake to generate file2.txt:

(base) user@comp:~$ snakemake file2.txt

It will see that rule two needs to be run, takes a look at the required input for this rule, and checks whether file1.txt already exists. If it does, it will immediatly execute rule two, if file1.txt does not already exist it will execute rule one first.

We highly recommend everyone interested in automating a part of their analysis to take a look at snakemake! For a more complete explanation of how snakemake works see the snakemake docs.