Revising trycycler and select assembly implementations

[fredjaya]

Tested and works on all Vibrio and Tenacibaculum barcodes. Currently on /scratch/er01/fj9712/2411_wholetest - to be moved.

Things to discuss and address either in this PR or later ones:

1. How should the results directory be structured? What files should be output here? (e.g. Medaka output for consensus chromosomes is split into 2 subdirectories #28)
2. Should the select assembly stats be reported?
3. What should go in the MultiQC report?
4. Does it work on the cat and dog data? (Fred currently testing)

Implementation

Reference-free chromosome assembly selection

Addresses #54, #23

For the chromosomal assembly, every barcode is assembled by flye and unicycler, and polished. The single "best" polished assembly out of flye, unicycler, and optionally trycycler (consensus assembly), is selected for downstream annotation and analyses.

To avoid biasing assemblies to published references, the assembly with the most complete BUSCOs is considered the best one. This now allows unicycler assemblies to be considered too. QUAST is also run but not used for selecting the best assembly.

Now only has a implementation for chromosomal assembly, instead of two independent ones, to make updating the criteria for selecting an assembly easier. For example, to incorporate QUAST outputs, or add additional tools like Merqury.

Trycycler implementation

Addresses #43, #60

Trycycler processes are now self-contained. Additional assemblers can be implemented easier to generate better consensus assemblies if required.

Added more error handling for too-few-contigs (trycycler cluster filters more out). If trycycler correctly fails at any point, the pipeline will still continue and select either the flye or unicycler assembly for downstream processes.

Input/output process definitions are more explicit (i.e. specific files instead of globs) for better error handling. A lot more operators and groovy in the workflow scope as a result.

[fredjaya]

This is what the current results/ folder looks like for a single barcode:

results
├── annotations
│   ├── barcode01
│   │   ├── abricate
│   │   │   └── barcode01_consensus_chr.txt
│   │   ├── amrfinderplus
│   │   │   └── barcode01_consensus_chr.tsv
│   │   ├── bakta
│   │   │   ├── barcode01_consensus_chr.faa
│   │   │   └── barcode01_consensus_chr.txt
│   │   └── plasmids
│   │       └── barcode01_bakta
├── assemblies
│   ├── barcode01_consensus
│   │   └── consensus.fasta
│   ├── barcode01_flye
│   │   └── consensus.fasta
│   ├── barcode01_plassembler
│   │   ├── flye_output
│   │   ├── logs
│   │   ├── plassembler_1730446699.3410256.log
│   │   ├── plassembler_plasmids.fasta
│   │   ├── plassembler_plasmids.gfa
│   │   ├── plassembler_summary.tsv
│   │   └── unicycler_output
│   ├── barcode01_unicycler
│   │   └── consensus.fasta
├── quality_control
│   ├── barcode01
│   │   ├── barcode01_consensus
│   │   ├── barcode01_consensus_busco
│   │   ├── barcode01_consensus.tsv
│   │   ├── barcode01_flye
│   │   ├── barcode01_flye_busco
│   │   ├── barcode01_flye.tsv
│   │   ├── barcode01_unicycler
│   │   ├── barcode01_unicycler_busco
│   │   └── barcode01_unicycler.tsv
│   ├── barcode01_kraken2
│   │   └── barcode01.k2report
├── report
│   ├── barcode01_consensus
│   ├── barcode01_flye
│   ├── barcode01_unicycler
├── run_info
│   ├── dag.svg
│   ├── gadi-nf-core-trace-*.txt
│   ├── report.html
│   └── timeline.html
├── taxonomy
│   ├── abricate_vfdb_output.txt
│   ├── amrfinderplus_output.txt
│   ├── barcode_species_table_mqc.txt
│   ├── combined_plot_mqc.png
│   └── phylogeny
└── tree

350 directories, 234 files

Suggestions:

• publish only the selected chromosomal assembly, remove assembler from name
• ...