Table 3. Details of the stochastic simulations performed to analyse the impact of the CTCF binding site, isochores, cryptic RSSs, and DNase hypersensitive sites on the edge distribution.
Formula | Iterations | Estimate | Std. Error | p-value |
---|---|---|---|---|
edges + nodecov("dnase") | 100 | 0.86291 | 0.07961 | <1e-04 |
edges + nodecov("ctcf") | 100 | 0.52386 | 0.04158 | <1e-04 |
edges + nodecov("rss") | 100 | 0.39780 | 0.03176 | <1e-04 |
edges + nodecov("iso") | 100 | -0.84035 | 0.09269 | <1e-04 |
As the summary table shows, after 100 iterations the statistics achieved a high degree of confidence (P-value < 0.0001). In particular, data demonstrate that DNase hypersensitive sites and CTCF binding sites have a positive influence on the presence of edges (nodecov.dnase = 0.86291 ± 0.07961 and nodecov.ctcf = 0.52386 ± 0.04158); at the same way cryptic RSSs are positively correlated with edges (nodecov.rss = 0.39780 ± 0.03176) although with less impact; on the other hand, isochores are negatively correlated with the edge distribution (nodecov.iso = -0.84035 ± 0.09269).