Table 4 2010-based predictions ranked by F*.

From: SANA: cross-species prediction of Gene Ontology GO annotations via topological network alignment

Rank

F*

NAF

Pair

M

EvCode

\(| {P}_{12}\cap {\Gamma }_{2}^{\prime}|\)

pred

Valid

Precision

1

0.500

2%

CE-HS

S3

IPI

3443

3131

1643

52.5%

2

0.496

2%

CE-HS

Import.

IPI

3443

3146

1634

51.9%

3

0.424

16%

SP-AT

Import.

IPI

455

569

217

38.1%

4

0.417

16%

SP-AT

S3

IPI

455

577

215

37.3%

5

0.416

5%

DM-HS

S3

IPI

4080

3237

1521

47.0%

6

0.404

5%

DM-HS

EC

IPI

5978

3204

1855

57.9%

7

0.375

8%

SP-AT

EC

IPI

472

968

270

27.9%

8

0.358

2%

CE-HS

EC

IPI

3443

2127

998

46.9%

9

0.346

2%

AT-HS

S3

IPI

5038

4360

1627

37.3%

10

0.341

2%

AT-HS

Import.

IPI

5038

4365

1605

36.8%

11

0.340

8%

MM-AT

S3

IPI

725

768

254

33.1%

12

0.309

5%

DM-HS

Import.

IPI

9018

3245

1895

58.4%

13

0.298

2%

SP-HS

S3

IPI

5380

4022

1401

34.8%

14

0.298

2%

SP-HS

Import.

IPI

5380

3985

1394

35.0%

15

0.291

5%

MM-AT

EC

IPI

755

1043

262

25.1%

16

0.277

1%

CE-HS

lgraal

IPI

3140

1511

644

42.6%

17

0.260

8%

SP-AT

graphlet

IPI

455

808

164

20.3%

18

0.259

8%

SP-AT

lgraal

IPI

472

827

168

20.3%

19

0.257

1%

SC-HS

EC

EXP

2111

1158

420

36.3%

20

0.249

2%

AT-HS

EC

IPI

5038

3629

1080

29.8%

  1. NAF: threshold that achieved F*; pair: species pair (cf. Table 1 (top)); M: topological measure; EvCode: evidence code supporting the annotation of the source (non-human) protein that produced the predicted human protein annotation; \(| {P}_{12}\cap {\Gamma }_{2}^{\prime}|\): intersection of the number of predictable in principle annotations (P12, see “Methods”) with \({\Gamma }_{2}^{\prime}\)—all annotations available in the validation set. pred: number of predicted annotations made using the specified source evidence code at the specified NAF (note this number can be bigger than the previous column since, clearly, any number of predictions can be made that may not appear in the validating GO release). valid: the number of validated predictions by any non-sequence-based evidence code. Precision: number of predictions that were validated.