Designing specific scoring file

[allind]

Thanks for this fabulous pipeline. It's a joy to use!

I have a specific use case that I'm struggling to code in the scoring file. I want to predict genes that may not have an encoded stop codon, but prioritize gene models that have a stop codon. In other words, if it looks like a gene except it lacks the stop codon, you retain the gene model, but if there is a similar gene model that does have a stop codon, that model is scored higher. I have figured out how to remove the stop codon as a requirement and get stop codonless genes, but I haven't been successful in getting Mikado to then prioritize gene models with a stop codon. Is this something that's possible? Thanks for your help.

[swarbred]

Sorry @allind missed this query

You would need to remove “is_complete” from the requirements sections to allow models without stop codons to be output. i.e. the expression under requirements: and as_requirements:

You can retain is_complete in the scoring section if you want to favour those models that do have stops. E.g.

scoring:
is_complete: {rescaling: target, value: true, multiplier: 3}

See
https://mikado.readthedocs.io/en/stable/Scoring_files/

The metrics section of that page gives all available metrics mikado calculates

https://mikado.readthedocs.io/en/stable/Scoring_files/#metrics

has_start_codon and has_stop_codon are both available so has_start_codon can be used to require or to score more highly those models with a start codon.

You may want to review https://mikado.readthedocs.io/en/stable/Tutorial/Scoring_tutorial/