The table of Subboundary Analysis output is shown automatically once you generate the analysis. If you closed it and wish to reopen the table, go to the Projects menu, choose table and select the "local leukemia boundaries: subboundary analysis results" table. Alternatively, right click on that analysis in the Results tab of the project window.
The table includes values and p-values for each of the subgraph statistics. The histograms show the distribution of each statistic, with the original statistic for the boundaries shown as a slim red bar.
Four of seven of the measures of "boundary-ness" were significant for the leukemia dataset. In essence, there were longer boundaries (higher length, fewer singletons, long path length between pairs of boundary elements) than expected by chance.
|
Statistic |
Meaning |
Value |
interpretation |
|
NS |
number of subboundaries |
significantly low |
boundary-like |
|
N1 |
number of singleton Boundary Elements |
significantly low |
boundary-like |
|
Lmax |
maximum subboundary length (number of linked BEs) |
significantly high |
boundary-like |
|
Lmean |
mean subboundary length |
not significant |
|
|
Dmax |
maximum subboundary diameter |
not significant |
|
|
Dmean |
mean subboundary diameter |
significantly high |
boundary-like |
|
D/L |
mean diameter-to-length ratio (indicates branchiness) |
not significant |
|
Subboundary diameter is the shortest path length between each pair of BEs in a subboundary.
So, there are significant areas of local change in leukemia. The next question is: are they associated with TCE injection wells?