Skip to main content

Table 2 Likelihood ratio tests for spatial clustering of CDK consensus matches

From: Clustering of phosphorylation site recognition motifs can be exploited to predict the targets of cyclin-dependent kinase

 

Number

H 0 (f s , f w )

H 1 (f1s, f1w) (f2s, f2w)

Λ

P value

'Known'

12

6.72, 10.8

25.1, 34.3; 2.66, 5.66

44.4

1.2 × 10-9

'Unbiased positives'

18

2.81, 8.51

19.8, 31.6; 1.53, 6.77

20.2

1.6 × 10-4

'Unbiased negatives'

173

0.67, 6.68

2.93, 47.7; 0.65, 6.34

5.58

0.13

'Known,' scrambled

12

0.96, 6.04

4.60, 10.2; 0.00, 4.48

5.21

0.15

  1. Comparison of a one-component versus two-component mixture of multivariate geometric distributions in different protein sets. Maximum likelihood parameter estimates (in matches per 1,000 residues) under the two hypotheses are indicated by f. See text for descriptions of parameters. Λ indicates the likelihood ratio test statistic, which is expected to be χ2 distributed with three degrees of freedom. P values are computed under that assumption. Seven low-confidence open reading frames were removed from the 'unbiased negatives', although similar results are obtained if they are included. CDK, cyclin-dependent kinase.
Back to article page

Genome Biology

ISSN: 1474-760X

Contact us