Data formats for modelling the spatial structure of chromatin based on experimental positions of nucleosomes

Abstract

In the nucleus of eukaryotic cells, DNA is wrapped around histone proteins, forming units termed nucleosomes. Nucleosome chains fold into chromatin. Despite extensive experimental advancement, many fundamental features of chromatin remain uncertain. Since all cell types and states cannot be profiled experimentally, especially in solution and in vivo, computer simulations are valuable tools for research. Most computer simulation models of chromatin are coarse-grained and describe the main characteristics of 3D chromatin packing. Newer models include experimentally derived positions of nucleosomes. While it is common practice in other disciplines, such as systems biology, to make experimental data publicly available, data from computer simulations of chromatin models are not usually published. Thus, data standard exchange formats are lacking, and we address this issue in the present work. We analysed the workflow, from experimental determination of the positions of nucleosomes through to analysis of outputs from simulated computer models. We defined standardized formats based on Extensible Markup Language (XML) for artefacts generated by steps in this workflow. We found that XML is useful since it is easy to transform XML-based-files by applying Extensible Stylesheet Language Transformations (XSLT) to other formats. We demonstrate the viability of this approach and the associated file formats using a complete example of computer simulation of chromatin domains based on experimentally determined nucleosome positions. The XML schemas and examples are published in an open source repository.