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
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
Milestone: generate negative data
Milestone: Feature selection
Milestone: ML-method selection
Milestone: Test different models, features, positive and negative data combinations
Milestone: Validation set
Milestone: Make a functional tool
Milestone: pick a jornal write paper