Statistics of genome architecture

• The mode of chromatin folding is termed genome architecture.
• The problem is reduced to the statistical analysis of length distribution.
• Empirical data can be approximated by De Finetti, gamma- or log-normal distribution.
• The log-normal distribution may reflect hierarchical chromatin folding.
• Structural entropy characterizes quantitatively inhomogeneity of architecture.

The main statistical distributions applicable to the analysis of genome architecture and genome tracks are briefly discussed and critically assessed. Although the observed features in distributions of element lengths can be equally well fitted by the different statistical approximations, the interpretation of observed regularities may strongly depend on the chosen scheme. We discuss the possible evolution scenarios and describe the main characteristics obtained with different distributions. The expression for the assessment of levels in hierarchical chromatin folding is derived and the quantitative measure of genome architecture inhomogeneity is suggested. This theory provides the ground for the regular statistical study of genome architecture and genome tracks.