AREM takes the classic peak calling algorithm used in MACS 1.4 (including an empirically optimal tag shift and a locally-adjusted poissonian statistical model for tag enrichment) and adds a component for realigning ambigously mapping reads. A probabilistic model assigns reads to regions that are more likely to be peaks (having higher enrichment).

For additional information, see our paper in RECOMB 2011 (D Newkirk, J Biesinger, A Chon, K Yokomori, X Xie, 2011) or download a copy here.

Regular releases of the software are pushed to the sourceforge page: http://sourceforge.net/projects/arem/

The development version of the code is available at: https://github.com/jakebiesinger/AREM