Roadmap

[teresa-m]

Roadmap

Implementation of an RNA-RNA interaction (RRI) prediction evaluation tool. The tools will take an RRI prediction of an RRI prediction tool and will evaluate how biologically relevant the prediction is. This evaluation will be base on an ML-model that is trained with RNA-seq interactome data. We will generate different models by splitting data into different ncRNA families or using different backgrounds.

What do we need to do?

This roadmap should help us to manage the tasks of the project.

Milestone: get data

• RNA interactome data from the following publication/Protocols: PARIS, SPLASH, and Liga-Seq
• find bacterial interactome data (later)

Milestone: generate negative data

• shuffling method
• context extension
• filter for not ML binders

Milestone: Feature selection

Milestone: ML-method selection

Milestone: Test different models, features, positive and negative data combinations

• find a validation measurement
• list all possible combinations
• get BW cloud running
• automate the run of different combinations on the BW cloud

Milestone: Validation set

• CopomuS data as a valiadtion set
• Find human or mouse validation set

Milestone: Make a functional tool

Milestone: pick a jornal write paper