How to use the international Vertebrate Genome Project workflows in Galaxy Australia

Anna Syme

What are the VGP workflows?

• These are workflows for genome assembly, developed for the Vertebrate Genome Project
• Website: https://vertebrategenomesproject.org/
• Data at Genome Ark https://vgp.github.io/genomeark/
• Paper: Rhie, A., McCarthy, S.A., Fedrigo, O. et al. Towards complete and error-free genome assemblies of all vertebrate species. Nature 592, 737–746 (2021).
• This paper covers the testing of tools and workflows. We recommend also looking at the Supplementary Information which is very informative.
• Most recent paper: Larivière, D., Abueg, L., Brajuka, N. et al. Scalable, accessible and reproducible reference genome assembly and evaluation in Galaxy. Nat Biotechnol (2024).

Are the VGP workflows in Galaxy Australia?

• An international team is working to create these workflows in Galaxy: for more information see https://galaxyproject.org/projects/vgp/
• In Galaxy Australia, you can access a set of these workflows by going to the Genome Lab, scroll to the Genome Assembly section, click on the Workflows tab.
• The workflows to import are:
  • Assembly with PacBio HiFi data:
    • BAM to FASTQ + QC v1.0
  • Assembly with PacBio Hifi and HiC data:
    • Kmer profiling
    • Hifi assembly and HiC phasing
    • HiC scaffolding
    • Decontamination

Can I use the VGP workflows in Galaxy Australia?

• Yes. You can run these on test data or real data. They are designed to work for vertebrate genomes where you have PacBio Hifi data and HiC data.
• The workflows are described in Galaxy Training Network materials.
• VGP assembly pipeline - short version, although still with complete workflow
• VGP assembly pipeline - long version

What data do I need?

Overall, you need these inputs:

• HiFi reads as collection
  • If HiFi data in BAM format, convert to FASTQ using the BAM to FASTQ + QC v1.0 workflow
  • then group output FASTQ files into a single collection
• HiC reads as R1 and R2

For each of the other workflows, you will need these inputs:

• WF1 Kmer profiling
  • Inputs:
  • HiFi reads in collection
• WF4 Hifi assembly and HiC phasing
  • Inputs:
  • HiFi reads in collection
  • HiC R1, HiC R2
  • from WF1: genomescope model parameters, genomescope summary, mery db
• WF8 HiC scaffolding
  • Inputs:
  • From WF4: Assembly in gfa format
  • From WF4: estimated genome size
  • HiC R1, HiC R2
  • Settings:
  • For restriction enzymes: set correctly
• WF9 Decontamination
  • Inputs:
  • From WF8: Scaffolded assembly in FASTA format
  • Settings:
  • For "Database for Kraken2": choose: Prebuilt Refseq indexes: PlusPF (Standard plus protozoa and fungi)
  • For "blast mitochondria db": select database/sequences: Locally installed Blast database: click the pencil icon: click in the text box to reveal a list of options: select RefSeq mitochondrion

Do I need any background knowledge before I run these workflows?

• We recommend the Galaxy Training Network tutorials.
• Introduction to Galaxy
• Assembly
• QC
• VGP tutorials
• These should be enough to get you started running the VGP workflows, but there are many additional tutorials that would also be useful.
• As most species have never had their genome sequenced, it is not possible to guarantee existing workflows are optimal for new data.
• It is most likely that any new genome assembly will have its own set of required workflow and analysis customisations to account for things such as ploidy and repeats.
• Usually, an assembly workflow will need testing and customising, in concert with reading the biological domain literature.

What is going on in the workflows?

• The workflows are large and have many steps.
• To better understand the workflow, look at the workflow canvas, view each tool and its settings, see how each step is connected.
• Consider what outputs you would require and make sure it is clear where they are and what they are called.
• For a description of many of the tools and the default parameters that have been set, see the Galaxy tutorial for VGP workflows.

Run the workflows on test data

• Look at the outputs and see if it makes sense and all tools ran.

Import real data

• If you will be using real vertebrate genome data, it is likely you will need more Galaxy storage. Contact the Galaxy Australia team to discuss/request.
• Import your real data sets into Galaxy.
• Or, to use real VGP data, go to Upload Data: Choose remote files: Genome Ark: species and choose a species.
• Note: not all species have data from all Hifi and HiC sources.
• Note that you will likely have to convert the data into the correct formats required. Alternatively, modify the workflows themselves to accept the data in the formats you have.

Is there anything I need to change in the workflows?

• Most likely, yes. Look at every tool and the parameters to check it suits your data and aims.
• Examples of things that might need changing:
• Set the kmer size in the meryl tool, and make sure the setting in the Genomescope tool matches this. There is no absolute answer for what this setting should be; a common setting is 21.
• In the Quast tool (used several times), set the lineage appropriately (e.g. eukaryote) and set as "large genome"
• In the Busco tool (used several times), set the lineage appropriately (e.g. Vertebrata)

Do I need to keep all of the output files from the workflow?

• No. In particular, some of the intermediate output files are very large, e.g., BAM files. You may wish to not keep these.
• To automatically not keep these files during the workflow run: go to the workflow canvas, see the box for a tool, see the output files, and untick any files you will not need.

Run the workflows on real data

• Modify workflows as needed: label or tag outputs, change tool parameters, swap tools, add or delete steps.
• For large data we would recommend testing tools and workflows on a subset of the data first.

Where to see the outputs

• Your Galaxy history holds the outputs from the workflow run.
• Some files may be hidden - click on "show hidden" to see.
• In the top Galaxy menu, go to User: Workflow invocations to see details of the workflow run.

What to do if a tool or workflow fails?

• Read the error message to see if you can troubleshoot the issue. Otherwise, contact help@genome.edu.au
• If you are able to, re-run the tool or workflow in case it was a temporary connection issue.