DIVAN


Welcome to the home page of DIVAN !

DIVAN (Disease-specific Variant ANnotation), is a feature selection, ensemble-learning framework for disease-specific noncoding variant annotation and prioritization. DIVAN considers thousands of epigenomic annotations and is able to handle the class imbalance and "large p, small n" problem. Unlike most existing computational tools, DIVAN is able to provide scores that gauge a variant’s impact in a disease-specific manner. Currently, DIVAN has been trained to evaluate variants for 45 different diseases/traits.

We have pre-computed the scores for each of the 45 disease/trait for each base of the whole genome (hg19). We also provide different software to score known variants and arbitrary genomic regions.

For questions or concerns, please contact Steve Qin (zhaohui.qin@emory.edu) or Li Chen (li.chen@emory.edu).


DIVAN is free and open-source software, released under the GNU General Public License v3.