Include Single Sample Predictor models in the StringTie pipeline

[Nicklas Nordborg]

A plug-in for running Single Sample Predictor (SSP) models was implemented in #1186 but it was postponed (see [5737]). It is time to include the SSP and run some models after StringTie has been completed.

The SSP plug-in need to be re-activated and we need to add a manual wizard for starting jobs and also include this in the auto-confirmation support.

[Nicklas Nordborg]

In 5920:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Re-enabled the SSP plug-in.

[Nicklas Nordborg]

In 5921:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Modified some descriptions about the plug-in and parameters and an annoying "null" error when none of the selected raw bioassays was StringTie bioassays.

[Nicklas Nordborg]

In 5922:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Added an items list for holding StringTie raw bioassays that should be analysed with the Single Sample Predictor plug-in.

Auto-confirmation will now start the SSP plug-in with all configured models.

Added a wizard to the Reggie start page for manually starting SSP. The wizard has not been implemented.

[Nicklas Nordborg]

In 5923:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

The manual StringTie confirmation wizard now add items to the Single Sample Predictor item list.

[Nicklas Nordborg]

In 5924:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Implemented a wizard for starting SSP manually.

[Nicklas Nordborg]

In 5925:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Improve error messages from plug-in by keeping track of the last exception and including that in the error message.

[Nicklas Nordborg]

In 5926:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Sort SSP models by name.

[Nicklas Nordborg]

In 5927:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Added possibility to include a description for a model.

[Nicklas Nordborg]

In 5929:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Added support for translating results from the SSP models. The main reason is that some models generate swedish text but we would like to keep annotations in english. Currently there are two translations:

• positiv -> Positive
• negativ -> Negative

[Nicklas Nordborg]

In 5937:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Implemented support for a "Scores" annotation type that is paired with the result annotation type for each SSP. The "Scores" annotation will contain a semicolon-separated list with scores for all possible classes. Each score is given as a key=value entry in the list. For example LumA=0.001; LumB=0.99; ...and so on.

[Nicklas Nordborg]

In 5941:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Added some checks for configured models before they are used:

• That an annotation type exists
• That the .Rdata file exists

If a check fails a warning messge is displayed and the model can't be used.

[Nicklas Nordborg]

In 5942:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Added checks before using configured models also in the auto-confirmation path.

[Nicklas Nordborg]

In 5943:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Increase the max number of raw bioassays to 500.

[Nicklas Nordborg]

In 5944:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Added a "Preset" selection list for selecting which models to run. The selection list is populated with existing plug-in configurations for the Single Sample Predictor plug-in.

[Nicklas Nordborg]

In 5945:

References #1239: Include Single Sample Predictor models in the StringTie pipeline

Added "Hög"->"High" and "Låg"->"Low" to translatations to make it compatible with the the pilot report.