Machine Learning, Bioinformatics and Geospatial Technologies
We develop predictive models on complex spatio-temporal patterns, integrating the different molecular biology levels, spatially explicit population models and landscape environmental analysis. We aim to create novel mathematical methods and ICT platforms connecting physiopathological patterns of disease with high dimensional data now available for Functional Genomics (e.g DNA microarrays, SNPs, proteomics, Deep Sequencing), with spatially epidemics simulation systems and geodatabases of environmental factors and socio-demographic data.
Our research is interdisciplinary. Since 1996, we have been working on biological and environmental data, porting novel Statistical Machine Learning Methods (classifiers and regression models) within Geographical Information Systems (GIS), with results in landscape epidemiology and environmental risk analysis. More recently, we started developing Individual Based Model (IBM) simulators of global pandemic scenarios (FLUMODCONT, EPIWORK). Since 2002, we also develop predictive machine learning methods for functional genomics: we design algorithms for predictive classification and profiling of high-throughput data, with implementation on standard workstations, computer clusters and in grid.
For our projects, we develop new software infrastructures (such as MITRIS) for data collection, management and analysis, supporting research and public agencies. We have created innovation in GIS (GRASS) and internet GIS (WebGIS) systems, and the spinoff company MPA Solutions in 2004. For high-throughput genomics data, we develop the Python/C machine learning package MLPY, several bioinformatics solutions, and we work for the integration of different genomics and patient data.
Finally, we actively promote the dissemination of interdisciplinary science with the WebValley Summer School Project, and by acting as a training lab for undergraduate and graduate students. We are proud to support projects of social interest that require visualizing and understanding data patterns such as the Intersos WebGIS for Humanitarian Response for Darfur and Tchad and the OECD Global Project on Measuring the Process of Societies. Since 2011, WebValley is international, with the participation of finalists of the INTEL ISEF Science Fair.
July, 2013: 1 grant funded by Bruno Kessler Foundation (FBK) on the topic "Predictive networks methods for functional metagenomics"
The candidate will develop original research on the integration of machine learning and network medicine methods within a state-of-art bioinformatics framework for the quantitative analysis of microbial communities of clinical interest. In particular, network-based methods for functional metagenomics studies will be applied to Next Generation Sequencing (NGS).
EVENT: iDSDn2010 - Next-generation Sequencing for Biomedical Omics 2010 (Castel Ivano, 20-21 Sept 2010), with CiBIO UniTN and IST Genova.
PAPERS: August 2010, the MAQC Consortium papers are out: with a main paper on Nature Biotechnology and Special Issue on The Pharmacogenomics Journal: see
- the FBK Press release (in Italian), also on the Repubblica/Le Scienze website here (in Italian).
- Nat Bio The MicroArray Quality Control (MAQC)-II study of common practices for the development and validation of microarray-based predictive models, Leming Shi and many authors from the MAQC-II Consortium (including, from FBK, C. Furlanello, S. Riccadonna, G. Jurman, R. Visintainer).
- All papers in The Pharmacogenomics Journal with MPBA authors are listed here.
- See in particular: An interactive effect of batch size and composition contributes to discordant results in GWAS with the CHIAMO genotyping algorithm, M Chierici, K Miclaus, S Vega, C Furlanello, The Pharmacogenomics Journal 10, 355-363
Special Website Section: ICT 4 Climate Change. (June 2010).
Other 2010 EVENTS:, "WPS and scientific computing for climate change: informal day" ENVIROCHANGE Workshop. Joint FBK & Cosbi "Informal Network Day" Workshop
PAPER: Paper and Editorial on February 2009 issue of Nature Genetics
Repeatability of published microarray gene expression analyses
John P A Ioannidis, David B Allison, Catherine A Ball, Issa Coulibaly, Xiangqin Cui, AedÃn C Culhane, Mario Falchi, Cesare Furlanello, Laurence Game, Giuseppe Jurman, Jon Mangion, Tapan Mehta, Michael Nitzberg, Grier P Page, Enrico Petretto & Vera van Noort
Mostly, your results matter to others -
High-throughput datasets and analysis protocols are intrinsically difficult to referee. Community standards enforced by journals may be less effective than is widely appreciated. Greater awareness of the needs and value of secondary data users can result in higher-impact papers.
Riva del Garda - Cortile della Rocca
Mercoledi' 3 Settembre 2008, ore 20:45