---
title: "nf-core/genomeassembler input discussion"
tags: pipeline,genomeassembler,planning,plans
---

# nf-core/genomeassembler input format discussion

Genome assembly can be a complex process with
a variety of data inputs and steps.

We'd like to formalise how the workflow takes input here:

## TSV format

Aimed at taking simpler input for less complex workflow execution:

Proposed properties:
- Should omit any one-to-many or many-to-one relations
- Could accept comma separated lists though where neccessary
- One row per assembly
- Column names should be identical to YAML format

## YAML format

Designed for more complex workflow executions.

Scope:
- Multiple samples
- Assembler execution with varying parameters
- What to subassemble/circularise?

samples:
  - id: My_awesome_species_1
    assemblies:
      - id: assemblerX_build1
        path: /path/to/assembly
      - id: assemblerY_build1
        path: /path/to/assembly
    hic:
      - id: hic_assembly_1
        path: /path/to/reads
    hifi:
      - path: /path/to/reads
      - path: /path/to/reads
    rnaseq:
      - path: /path/to/reads
    isoseq:
      - path: /path/to/reads
tools:
  assembler1:
    options:
      - '--opt1 X --opt2 Y'
      - '--opt2 A --opt2 B --opt3 C'
  assembler2:
    options:
      - '--optx X --opty Y'
      - '--optx A --opty B --optz C'
  assembler3: true

- Should use [Pep Specification](http://pep.databio.org/en/latest/).

### Implementation details.

- Tool options are configured using the `ext.args` process option. In order to use this in the configuration, the options must be passed with the `meta` input map/dictionary, and can then be accessed in the configuration using:

  process {
      withName: 'ASSEMBLER_X' {
          ext.args = { meta.assembler_options }
      }
  }