Home Research
Home Research

Prior Knowledge Integration in high-dimensional molecular data

Research Area: Computational Biology and Biostatistics
Status: In progress  
Faculty: Annalisa Barla Participants: Grzegorz Zycinski, Margherita Squillario


Modern biology is facing the major issue of integrating well-established statistical analysis approaches with the available knowledge stored in public databases and literature, in order to obtain interpretable results that may be applied in actual biological practice.


The main focus is to take a step towards the integration of two different aspects of data mining: database and machine learning perspective.

We implemented KDVS, a framework for the analysis of high-throughput data that combines data and domain knowledge.

KDVS is flexible enough to use not only microarray data, but other high-throughput data of choice (e.g. from mass spectrometry, microarray, next generation sequencing). Moreover, the feature selection phase incorporates prior biological knowledge in a modular way from various repositories and is ready to host different statistical learning techniques.