Evaluation Results

6. arginine biosynthesis

Similarity values obtained by G-SESAME

	ARG1	ARG2	ARG3	ARG4	ARG5,6	ARG8	ECM40
ARG1		0.155	0.235	0.235	0.227	0.235	0.155
ARG2			0.218	0.128	0.176	0.218	0.932
ARG3				0.199	0.270	0.338	0.218
ARG4					0.193	0.199	0.128
ARG5,6						0.270	0.181
ARG8							0.218
ECM40

Clustering results obtained by G-SESAME

Threshold

Initial

0.930

0.330

0.270

0.230

0.210

Clustering
Result

ARG8

ARG3

ARG5,6

ARG4

ARG1

ECM40

ARG2

ARG8

ARG3

ARG5,6

ARG4

ARG1

ECM40
ARG2

ARG8
ARG3

ARG5,6

ARG4

ARG1

ECM40
ARG2

ARG8
ARG3
ARG5,6

ARG4

ARG1

ECM40
ARG2

ARG8
ARG3
ARG5,6
ARG4
ARG1

ECM40
ARG2

ARG8
ARG3
ARG5,6
ARG4
ARG1
ECM40
ARG2

Similarity values obtained by Resnik's method

	ARG1	ARG2	ARG3	ARG4	ARG5,6	ARG8	ECM40
ARG1		0.281	0.281	0.281	0.281	0.281	0.281
ARG2			1.378	0.281	1.013	1.378	5.755
ARG3				0.281	1.013	1.378	1.378
ARG4					0.281	0.281	0.281
ARG5,6						1.013	1.104
ARG8							1.378
ECM40

Clustering results obtained by Resnik's method

Threshold

Initial

5.755

1.355

1.055

0.255

Clustering
Result

ECM40

ARG2

ARG8

ARG3

ARG5,6

ARG4

ARG1

ECM40
ARG2

ARG8

ARG3

ARG5,6

ARG4

ARG1

ECM40
ARG2
ARG8
ARG3

ARG5,6

ARG4

ARG1

ECM40
ARG2
ARG8
ARG3
ARG5,6

ARG4

ARG1

ECM40
ARG2
ARG8
ARG3
ARG5,6
ARG4
ARG1

Annotation information for ARG3 and ARG5,6:

Red term annotates gene ARG3. Green terms annotate gene ARG5,6.

Annotation information for ARG3 and ARG2:

Red term annotates gene ARG3. Green term annotates gene GNA1.

Analysis:

The annotation information for ARG2, ARG3 and ARG5,6 in the above figures show that the distance between the GO terms annotating ARG3 and ARG5,6 respectively is 6. On the other hand, the distance between the GO terms annotating ARG3 and ARG2 respectively is 9.

Therefore, the similarity between ARG3 and ARG5,6 should be greater than the similarity between ARG3 and ARG2 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 ARG3 and ARG2 is greater than the similarity between ARG3 and ARG5,6, inconsistent with the annotation information.

Conclusion:

G-SESAME is better.