In this talk, I will give an overview of our research work at Tulane Multiscale Bioimaging and Bioinformatics Laboratory (http://www.tulane.edu/~wyp/
). First, I will present our latest developments of computational and statistical models such as sparse regressions for data representation and analysis. Second, I will present examples of applying these models to the extraction of biomarkers from (epi) genomics and MRI imaging data. Third, I will focus on the integration of multiscale genomic and imaging data for improved diagnosis of mental illnesses (e.g., schizophrenia). Finally, I will show how to build brain networks and use them as fingerprints to recognize different age groups and predict IQs.