Skip to main content

Table 2 X-CAP prioritizes causal stopgains in patient exomes. Each row in the table describes a single patient, the causative gene and variant, the genotype of the variant, and the percentile-normalized score provided by each classifier. For each method, raw scores were percentile-normalized in comparison to the scores output by the classifier on the test set of \(\mathcal {D}_{\text {original}}\). All ten patients contain one rare stopgain and no other rare mutations in the causal gene. violetBolded entries have the highest percentile for a given variant. redItalicized entries would have been misclassified on the basis of the original authors’ recommendations (CADD, DANN, and Eigen do not provide a decision rule). X-CAP assigns the highest percentile six out of the ten times and mischaracterizes only one variant. No other tool assigns the highest percentile-normalized score more than once, and MutPred-LoF and ALoFT mischaracterize variants five and three times, respectively

From: X-CAP improves pathogenicity prediction of stopgain variants

Patient ID

Gene

HGVS

GT

X-CAP

MutPred-LoF

ALoFT

CADD

DANN

Eigen

DDDP108441

FOXP1

c.C1366T:p.Q456X

0/1

89.4

87.7

89.4

89.2

81.3

violet95.1

DDDP108556

MED12

c.C5916A:p.Y1972X

0/1

90.8

red26.6

violet93.3

23.7

52.0

8.9

DDDP108105

SATB2

c.C1375T:p.R459X

0/1

violet98.8

77.3

96.4

57.2

64.1

70.0

DDDP109873

EP300

c.C5581T:p.Q1861X

0/1

90.8

violet98.9

98.8

57.2

44.1

46.5

DDDP111266

CASK

c.C613T:p.R205X

1/1

violet99.1

red17.9

97.5

46.0

81.3

6.8

DDDP107416

AUTS2

c.C976T:p.Q326X

0/1

red54.0

78.6

red67.4

17.9

violet81.3

45.3

DDDP108492

DYRK1A

c.C691T:p.R231X

0/1

violet98.7

red19.5

90.7

57.2

81.3

70.7

DDDP100091

KDM6A

c.C3047A:p.S1016X

0/1

violet93.6

red30.8

93.1

66.0

64.1

11.2

DDDP110976

POGZ

c.T2579A:p.L860X

0/1

violet93.3

red22.9

red79.0

12.3

14.4

25.2

DDDP110748

ANKRD11

c.C1801T:p.R601X

0/1

violet92.6

83.2

red69.5

10.0

21.0

15.2