Deliverables

Deliverable WP1. The report provides a summary of the scientific activities carried out for the Work Package WP1 that primarily focused on the problem of multi-omics unsupervised data integration in the context of low-rank dimensionality reduction techniques and non-negative matrix factorization (NMF) models and their constrained extensions, together with the joint approaches. Within this context, the UniBA research unit proposed the modified joint tri-nonnegative matrix factorization model (modjointNMF), that focuses on vertical integration of different omics datasets having the same rows n (the number of samples is assumed to be equal non each omics) and have been processed, normalized or scaled depending on the omics type. Additionally, they investigate other advanced methods for analyzing single-omics data (in particular, genomic data and differentially expressed genes).

Deliverable WP2. The report provides a summary of the scientific activities carried out for the Work Package WP2  that primarily focused on the problem of multi-omics supervised data integration in the context of survival analysis. Within this context, the IAC-CNR research unit first explored the state of the art in the literature. Then, they proposed AFTCoop, a novel cooperative survival analysis method within the accelerated failure time (AFT) model framework. Additionally, they investigate other advanced methods for analyzing single-omics data in the context of survival.

Deliverable WP3. The reports focused on implementing and benchmarking new algorithms and computational packages for multi-omics supervised data integration in survival analysis and novel algorithms for unsupervised multi-omics data integration with a specific focus on Non-negative matrix factorization approaches. These practical activities build on the methodological advances from WP1 and WP2. Specifically for WP2 two novel packages are made available as open-source tools on GitHub: the R package AFTCoop, and the R Package Omics2Surv. Additionally contributed to the development and maintenance of the other two packages: the R package AFTNet, and the R package INetTool are also considered.

Deliverable WP4. The report provides a summary of the scientific activities carried out for the Work Package WP4 that  focused on the application of models developed in WP1 and WP2 to a few case studies in cell development and cancer. The benchmarking of methods using data from the literature or simulated data are instead reported in the publications related to the project. The practical activities described are mainly build on the methodological advances performed on the basis of the theoretical results obtained in work-packages WP1 and WP2, togethr with some additional integrative pipelines to deal with real datasets from domain-specific examples such as a type of lymphoma, multiple myeloma and some specific lipidomic data. 

Deliverable WP5. The report outlines how project results have been shared with the scientific community and all the activities related to Work Package WP5, that was mainly devoted to disseminating the project results and transferring knowledge to students and young researchers from various scientific backgrounds. The disseminating activities have been mostly concentrated in the second year of the project after preliminary results related to work packages WP1 and WP2 where available. All the disseminating activies have been also reported in the Seminars and Talks, Meeting and Events and Publications related pages.  

To access the full documents, a request can be sent to the Principal Investigator at the following email address: nicoletta.delbuono@uniba.it