Table 3.

Summary of physical parameters obtained from the fit of Boettiger’s data [2] for active (A), inactive (I) and repressed (R) epigenetic domains through the Bayesian procedure (mean values, see Additional file 1: Figs. S3–S5)

State	Active		Inactive	Repressed
	Bayesian fit	Estimate	Bayesian fit	Bayesian fit
Fitted:
$\epsilon [k_{B}T]$	$0.1\pm 0.05$		$0.32\pm 0.03$	$0.44\pm 0.04$
$K_{\mathrm{bp}}$ [bp]	$K_{\mathrm{nm}}\propto K_{\mathrm{bp}}^{0.56}$	~ 1100–1500	$3900\pm 1300$	$1500\pm 550$
$K_{\mathrm{nm}}$ [nm]		~ 32–37	$60\pm 9$	$37\pm 6$
$a_0$ [nm]	$130\pm 7$		$93\pm 10$	$94\pm 4$
$a_{\infty }$ [nm]	$290\pm 15$		$170\pm 10$	n/a
$N_0$	$630\pm 370$		$10\pm 6$	n/a
Derived:
$c=\frac{K_{\mathrm{bp}}}{K_{\mathrm{nm}}}$ [bp/nm]		~ 35–40	$66\pm 24$	$40\pm 16$
$c_{10}$ [nucl./10 nm]		~ 1.9–2.2	$3.5\pm 1.5$	$2\pm 1$
C [$\text{nucl.}/K_{\mathrm{nm}}$]		~ 6–8	$20\pm 7$	$8\pm 3$

Errors are calculated from the standard deviations of marginalized parameter distributions. At the bottom, some derived geometrical parameters as the compaction in bp/nm, in nucleosomes per 10 nm, the number of nucleosomes per Kuhn segment C. Derived parameters are calculated by assuming a nucleosome repeat length of 182 bp for active domains, 192 bp for inactive and repressed domains [25]. (The numerical results obtained with 182 or 192 bp are very close, in the error range). For active domains, the right column estimates are obtained by including architectural features, see “Discussion”