Statistical modelling of multi-omics data
Development of statistical models for the prediction of proteomics data from transcriptomic data (or generally one type of data from another).
Key features of the developed models will be:
- High-dimensional inference approaches, that can deal with a large number of proteins as predictors and/or as response of the multivariate regression model
- Statistical distributions of outcomes beyond Gaussian, such as for the case of censured data.
- Novel penalty functions that can point to the identification of regulatory motifs or hotspots, and that go beyond the identification of single important predictors
- Software implementation
Bioscientists can benefit from this research.