Similarity values obtained by G-SESAME

  GCV2 GCV1 ADE3 MTD1 MIS1
GCV2  10.2490.2940.249
GCV1    0.2490.2940.249
ADE3      0.4621
MTD1        0.462
MIS1         

Clustering results obtained by G-SESAME

Threshold Initial1.0000.4600.290
Clustering
Result
MIS1

ADE3

MTD1

GCV1

GCV2

MIS1
ADE3

MTD1

GCV1
GCV2

MIS1
ADE3
MTD1

GCV1
GCV2

MIS1
ADE3
MTD1
GCV1
GCV2

Similarity values obtained by Resnik's method

  GCV2 GCV1 ADE3 MTD1 MIS1
GCV2   8.4620.8301.3780.830
GCV1     0.8301.3780.830
ADE3       3.6106.953
MTD1         3.610
MIS1          

Clustering results obtained by Resnik's method

Threshold Initial8.4636.8633.5631.363
Clustering
Result
MIS1

ADE3

MTD1

GCV1

GCV2

MIS1

ADE3

MTD1

GCV1
GCV2

MIS1
ADE3

MTD1

GCV1
GCV2

MIS1
ADE3
MTD1

GCV1
GCV2

MIS1
ADE3
MTD1
GCV1
GCV2

Analysis:

Genes ADE3 and MIS1 are both annotated by three molecular function terms: GO:0004329: formate-tetrahydrofolate ligase activity, GO:0004477: methenyltetrahydrofolate cyclohydrolase activity, and GO:0004488: methylenetetrahydrofolate dehydrogenase (NADP+) activity. GCV1 and GCV2 are annotated by the same molecular function term GO:0004375: glycine dehydrogenase (decarboxylating) activity.

Therefore, ADE3 and MIS1 have the same molecular function. GCV1 and GCV2 have the same molecular function.

The similarity values and clustering results obtained by G-SESAME are consistent with these observations.

However,based on Resnik's method, the similarity between GCV1 and GCV2 is greater than the similarity between ADE3 and MIS1. This is inconsistent with the human perception.

Conclusion:

G-SESAME is better.