Documentation for Snakemake workflow "genome hybrid assembly"

Disclaimer the workflow is in prototype state and configuration may change at any time

This workflow creates state-of-the-art genome hybrid assemblies for diploid vertebrate species. This version of the workflow was developed for the following scenario:

• input species: human
  • successfully tested as well: muntjac
• assembler: Verkko v1.4.1
  • hifiasm v0.19.x is not yet fully integrated
• inputs:
  • long accurate reads: PacBio HiFi (Sequel-II/Revio)
    • required coverage: at least ~40X, ideally ~60X
  • long connecting reads: Oxford Nanopore ultralong (R9)
    • required coverage: ~30X ultralong (>100 kbp) reads
  • optional input for phasing:
    • trio: kmer databases created with meryl
    • HiC: HiC short reads
    • graph/node coloring for Verkko's Rukki (GFA file)
• outputs:
  • main: whole-genome assembly, potentially phased
  • main: basic (length) statistics about assembly and long reads
  • optional: a coordinate map between the homopolymer-compressed assembly graph and the linearized plain FASTA files

The sample sheet must be a TAB-separated text file (.tsv file extension) with at least the columns sample, hifi and ont, where both hifi and ont columns can hold an arbitrary number of input file paths (comma-seperated, i.e., file_path1,file_path2,file_path3) representing the respective read dataset for that sample. Common file extensions are recognised (e.g., fastq.gz, .fq.gz and so on). The Verkko assembler can optionally be configured for using three different phasing signals; add the column target to the sample sheet plus the following fields:

1. trio-based: set value trio in column target and add columns hap1 and hap2 pointing to meryl k-mer databases of the sample parents (conventionally, hap1 should be the father and hap2 the mother)
2. Hi-C: set value hic in column target and add fields hic1 and hic2 for the Hi-C reads of mate 1 and 2, respectively
3. Strand-seq: set value sseq in column target and add field phasing_paths pointing to a .gaf format file produced by the Grapahasing pipeline

Since Verkko itself is implemented as a Snakemake workflow, you can execute a dry run to check if all input requirements are met by setting the option verkko_dry_run to true, see this example configuration:

Example parameterization

User documentation for workflow template

All standard workflows of the CUBI implement the same user interface (or at least aim for a highly similar interface). Hence, before executing the workflow, we strongly recommend reading through the documentation that explains how we help you to keep track of your analysis results; we refer to this concept as "file accounting". This feature of standard CUBI workflows enables the pipeline to auto- matically create a so-called "manifest" file for your analysis run.

In case of questions, please open a GitHub issue in the repository of the workflow you are trying to execute.

Developer documentation

Besides reading the user documentation, CUBI developers find more information regarding standadized workflow development in the developer notes. Please keep in mind to always cross-link that information with the guidelines published in the CUBI knowledge base.

Please raise any issues with these guidelines "close to the code", i.e., either open an issue in the knowledge base repo or in the affected repo for more specific cases.