Table 3 |
||||||||
|
Overview of the characteristics of the GI analyses using decreasing similarity thresholds (for all GIs > 10 kb) |
||||||||
|
Stringency |
Total number of GIs |
Number of genomes |
GI < CI |
Clusters |
GIs in clusters* |
Percentage clustered (%) |
Prediction Accuracy (%) |
|
|
|
||||||||
|
CI-0 |
++++ |
2191 |
267 |
1 |
20** |
40** |
1.8 |
99.9 |
|
CI-5 |
+++ |
2047 |
260 |
9 |
99 |
202 |
10.0 |
98.6 |
|
CI-10 |
++ |
1787 |
246 |
11 |
134 |
271 |
15.3 |
97.5 |
|
CI-25 |
+ |
1370 |
220 |
16 |
185 |
383 |
28.3 |
94.8 |
|
|
||||||||
|
Total analyzed |
2609 |
322 |
||||||
|
|
||||||||
|
The totals represent the total numbers in the original data set from IslandViewer. *) The percentage of clustered GIs (second last column) excludes 17 GIs from the total number of GIs (third column), since there are 17 genomes with a single GI only, and with less than two GIs there can be no clustering. **) Six out of 20 clusters contain in fact largely identical Genomic Islands, which explains their high compositional similarity. |
||||||||
|
Roos and van Passel BMC Genomics 2011 12:427 doi:10.1186/1471-2164-12-427 |
||||||||
