Neurodegenerative diseases (NDDs), such as Alzheimer's disease or Parkinson's disease, are defined by the progressive neuronal death in distinct brain areas. An important feature of this conditions is the heterogeneity of their symptoms driven by the complex interaction of environmental, molecular, and genetic factors.

This heterogeneity presents a challenge to understanding the mechanisms involved in their pathogenesis and progression. The increased number of available omics data, such as genomics or proteomics, has allowed researchers to identified specific genetic drivers of the disease. However, studies are usually only focused on one single omic type, i.e., a single level of biological information.

Here, we propose to develop new computational strategies using machine learning algorithms to integrate different omics studies to obtain a comprehensive biological model of NDDs. We will address current data management challenges such as missing values or high dimensionality by refining feature selection methods to reduce complexity and noise.

Then, we will develop a software to integrate these omics datasets, ultimately building an optimized and simpler representation of the data. With our validated model and developed software, we will apply this approach to understand disease mechanisms. Specifically, we expect to identify potential molecular mechanisms through all the layers of information and to create an omic pattern associated with a neurodegenerative disorder.

The execution of the present project will entail important benefits for the scientific community through the availability of the developed integrated data analysis tool and aiding in the discovery of key molecular drivers in NDDs, bringing us closer to targeted, effective treatments.