Modeling nucleosome position distributions from experimental nucleosome positioning maps

* ABSTRACT Motivation: Recent experimental advancements allow determining positions of nucleosomes for complete genomes. However, the re- sulting nucleosome occupancy maps are averages of heterogene- ous cell populations. Accordingly, they represent a snapshot of a dynamic ensemble at a single time point with an overlay of many configurations from different cells. To study the organization of nu- cleosomes along the genome and to understand the mechanisms of nucleosome translocation, it is necessary to retrieve features of specific conformations from the population average. Results: Here, we present a method for identifying non-overlapping nucleosome configurations that combines binary-variable analysis and a Monte Carlo approach with a simulated annealing scheme. In this manner we obtain specific nucleosome configurations and opti- mized solutions for the complex positioning patterns from experi- mental data. We apply the method to compare nucleosome position- ing at transcription factor binding sites in different mouse cell types. Our method can model nucleosome translocations at regulatory genomic elements and generate configurations for simulations of the spatial folding of the nucleosome chain. Availability: Source code, precompiled binaries, test data and a web-based test installation are freely available at http://bioinformatics.fh-stralsund.de/nucpos/.