My Research project at Massachusetts Institute of Technology and Harvard – ARTIFICIAL INTELLIgeneCE, is designed to sort the ‘best set of attributes’ out of massive data sets responsible for Bipolar disorder with an aim to improve the chances of prevention and cure of the disease. Project AI develops a new mode of attribute discovery and external cross validation methods using a unique, data-driven, integrative Bayesian approach to merge gene expression data from Bipolar disorder – related experiments into two prognostic models.

The designed model proved to be successful with an AUROC (Area under Receiver Operating Characteristic) of 0.907 on bipolar disorder samples finding six genes – C8ORF44, ADH5, MCL1, PDE1A, ASPH, NTM as responsible for the disease. Future studies can shed some light on these relationships and the functions of these genes and gene products, also in AIDS, Cancer.

External Crossvalidation

The Six genes responsible for Bipolar disorder

Plot of AUROC