Technical Report Details
| Date |
17-6-2010 |
| Number |
DISI-TR-10-02 |
| Title |
A computational approach for the identification of potential biomarkers on heterogeneous data |
| Authors |
Margherita Squillario, Annalisa Barla |
| Bibtex Entry |
DISI-TR-10-02 |
| E-mail |
margherita.squillario@unige.it |
| Link |
http://slipguru.disi.unige.it/Research/papers/TR-DISI-10-02.pdf |
| Abstract |
Background: Alzheimer’s Disease (AD) is the most spread form of dementia among the elderly population and until now it has no definitive cure. A reliable characterization at a molecular level of this disease is therefore necessary for an effective clinical test for early
diagnosis. In the context of molecular data analysis, we aim at using a recently proposed classification and feature selector method, namely l1 l2 regularization with double optimization, to assess the biological soundness of its results. Furthermore, we aim at verifying the
existence of both direct and functional overlap of different signatures from heterogeneous data (one protein abundance and two microarray datasets).
Results: The biological validity of the relevant genes identified by l1 l2 was assessed querying public repositories, such as the Gene Ontology, KEGG and Entrez. This certified biological knowledge revealed that some of the identified genes are already known to be involved in AD, some are known to be involved in other brain diseases or just to be expressed in the brains, while others could be interesting for further investigation. Moreover, the concerted analysis performed on the three signatures, lead to a functional ovelap both for the GO and the KEGG pathways.
Conclusions: The analysis of the three different datasets let us to identify signatures with a relevant functional overlap that highlights pathways and ontologies significant for AD. Additionally, the method identified non-overlapping genes, some of which already
known to be involved in the disease, confirming the necessity to use heterogeneous data to broaden the understanding of complex diseases. |
|
|
 |
