Used for FAIRE-seq. Usefull for noisy data, but a little bit complex to use, but still less complex than ZINBA. You have to generate background .bff files with the help of GEM tools and convert .mappability file into .wig file, but .iff files are more obscure. Fortunately, Furey's lab provided some generic ones. Nonetheless I never could reproduce exactly their peak calling using ENCODE data... it seems there is a part of randomness in the process. Also, note that you need an additionnal perl script to calculate peaks p-values !