The complete pipeline

The complete pipeline#

Let us build the complete pipeline, then alter it to demonstrate versatility.

Note

This pipeline uses workflow elements that were built with linux/macOS in mind, and as such may not run on Windows at this time.

Construct the workflow#

Open GRAPEVNE (or clear the graph view).

Select (vocpl) from the repository filter. Here you will find modules corresponding to the different stages of our pipeline. Following the workflow presented on the previous page, drag the following items into the graph-view and connect them together, taking care to connect them in the correct order using the In or fasta inputs (do not connect the seeds inputs). Use the fasttree module for maximum likelihood estimation.

Your workflow should now contain the following nodes:

  • subsample_alignment

  • nextalign

  • fasttree

  • treetime

  • dta

That’s our working pipeline complete! We now need to provide two pieces of information: a seeds.txt file, containing the seeds for our analysis, and the beta.fasta file, which is available through controlled access via GISAID (if you are attending a GRAPEVNE workshop then this file will be provided in-person).

In order to include a file (the beta.fasta file) from your local file system we have some choices, but the simplest is to drag the LoadFile module into our graph at the top of the workflow and connect it to the fasta input. Within this module, ensure that the path to the local file is correctly specified.

Finally, we need to provide the seeds for our analysis. We will do this by using a module that supplies a seeds-file, provide_seeds. Conceptually, this module could be replaced with a local file, a database query, or a prompt which asks the user which seeds to use. However, we are demonstrating that resources can be provided along with our workflow modules (in-fact, all scripts that are used in this analysis are also resources which are downloaded automatically when the workflow is run).

We now need to connect the provide_seeds module in to our workflow. But where should we connect it? While intuitively you might want to connect it to the first module (provide_seeds), or indeed connect it to all modules (which would also work), we actually only need to connect the provide_seeds module in to the final module of the workflow: the dta module in our case. This is because snakemake is a build-system, which means that we specify the desired output, and snakemake will work out the necessary steps to provide that result. Our graph acts to provide a clear sequence of steps that must be undertaken to take us from the provide beta.fasta file, to the desired trait anlaysis for our seeds of interest (dta and provide_seeds).

You workflow should now be ready to run. Select Build & Test and then wait for the workflow to finish (this should take less than 10 mins on a modern laptop).

Change the tree estimation method#

By now the workflow should have completed and you should have access to the various files generated during the run.

Let us now imagine that we want to change the workflow and try out a different maximum likelihood method for tree estimation. In particular, we want to replace the fasttree module with another: the iqtree module.

Delete the fasttree module. Drag in and connect up the iqtree module. Done.

Re-run the workflow.

Contents