Similarity values obtained by G-SESAME
| | HEM1 | HEM2 | HEM3 | HEM4 | HEM12 | HEM13 | HEM14 | HEM15 |
| HEM1 | | 0.128 | 0.264 | 0.128 | 0.128 | 0.128 | 0.128 | 0.193 |
| HEM2 | | | 0.235 | 0.728 | 0.338 | 0.199 | 0.199 | 0.482 |
| HEM3 | | | | 0.235 | 0.235 | 0.235 | 0.235 | 0.328 |
| HEM4 | | | | | 0.338 | 0.199 | 0.199 | 0.482 |
| HEM12 | | | | | | 0.199 | 0.199 | 0.482 |
| HEM13 | | | | | | | 0.728 | 0.284 |
| HEM14 | | | | | | | | 0.284 |
| HEM15 | | | | | | | | |
Clustering results obtained by G-SESAME
| Threshold | Initial | 0.720 | 0.480 | 0.320 | 0.280 | 0.260 |
Clustering
Result |
|
|
|
HEM4
HEM2
HEM15
HEM12
HEM3
|
|
HEM4
HEM2
HEM15
HEM12
HEM3
HEM14
HEM13
|
|
HEM4
HEM2
HEM15
HEM12
HEM3
HEM14
HEM13
HEM1
|
|
Similarity values obtained by Resnik's method
| | HEM1 | HEM2 | HEM3 | HEM4 | HEM12 | HEM13 | HEM14 | HEM15 |
| HEM1 | | 0.281 | 1.378 | 0.281 | 0.281 | 0.281 | 0.281 | 0.281 |
| HEM2 | | | 0.281 | 5.258 | 1.378 | 0.281 | 0.281 | 1.378 |
| HEM3 | | | | 0.281 | 0.281 | 0.281 | 0.281 | 0.281 |
| HEM4 | | | | | 1.378 | 0.281 | 0.281 | 1.378 |
| HEM12 | | | | | | 0.281 | 0.281 | 1.378 |
| HEM13 | | | | | | | 5.904 | 0.281 |
| HEM14 | | | | | | | | 0.281 |
| HEM15 | | | | | | | | |
Clustering results obtained by Resnik's method
| Threshold | Initial | 5.904 | 5.204 | 1.304 | 0.204 |
Clustering
Result |
|
|
|
|
HEM4
HEM2
HEM15
HEM12
HEM3
HEM1
HEM14
HEM13
|
|
Annotation information for HEM1 and HEM3:
Red term annotates gene HEM1. Green term annotates gene HEM3.
Annotation information for HEM12 and HEM15:
Red term annotates gene HEM12. Green term annotates gene HEM15.
Analysis:
The annotation information for HEM1, HEM3, HEM12 and HEM15 in the above figures show that the distance between the GO terms
annotating HEM1 and HEM3 respectively is 8 hops. On the other hand, the distance between the GO terms annotating HEM12 and HEM15
respectively is only 4 hops.
Therefore, the similarity between HEM12 and HEM15 should be greater than the similarity between HEM1 and HEM3 based on the
annotation information.
The similarity values and clustering results obtained by G-SESAME are consistent with this observation.
However, based on Resnik's method the similarity between HEM1 and HEM3 is equal to the
similarity between HEM12 and HEM15, inconsistent
with the annotation information.
Conclusion:
G-SESAME is better.