Table 1 Genes with elevated non-silent mutation rates.

From: A multigene mutation classification of 468 colorectal cancers reveals a prognostic role for APC

Gene

N †

AA length

Rate

Residual scores §

|| Codons ⩾8

¶ HRun

    

Overall

MSI

MSS

  

KRAS

192

189

72.4

13.52

3.46

14.09

4

0

TP53

292

393

52.9

16.25

3.67

17.1

6

0/1

HLA-A*

102

365

19.9

8.52

3.91

8.43

1

0

APC

494

2,843

12.4

17.08

—

18.75

9

15/18

SMAD4

59

552

7.6

5.16

—

5.68

1

0

FBXW7

62

707

6.2

4.92

3.12

4.54

2

0/1

MUC4*

463

5,412

6.1

13.30

4.72

14.06

5

0/3

BRAF

61

766

5.7

4.69

5.35

—

1

0/1

TCF7L2

48

602

5.7

4.16

—

4.32

 

11/32

PIK3CA

76

1,068

5.1

4.97

—

4.82

3

0

GNAS

62

1,037

4.3

4.09

—

3.77

 

0

TAF1L

70

1,826

2.7

3.09

—

—

 

0/9

CSMD3

115

3,707

2.2

3.06

—

—

 

0/6

LRP1B

138

4,599

2.1

3.19

—

4.02

 

1/3

OBSCN*

220

7,968

2.0

3.49

—

3.56

 

0/7

SYNE1*

227

8,797

1.8

3.08

—

4.01

 

0/1

TTN*

845

34,350

1.8

5.30

—

6.76

 

3/23

CBX4*

31

560

3.9

—

3.10

—

1

0/5

ITGB4*

54

1,822

2.1

—

3.71

—

 

0/4

ADAMTS18

52

1,210

3.1

—

—

3.08

 

0/19

FAM123B

41

1,135

2.6

—

—

3.13

 

0/5

MUC16*

295

14,507

1.4

—

—

3.78

 

0/3

  1. AA, amino acid; GATK, Genome Analysis ToolKit; MSI, microsatellite instability; TCGA, The Cancer Genome Atlas.
  2. †N=number of non-silent mutations. Rate=105 × N/(468(3 × AA length)). §Residual scores denote the z-score for difference of the mutation count from the fitted robust regression line (SAS 9.3). Residual scores ⩾3 are considered to be outliers. A larger score represents a more substantial difference. ||Codons ⩾8=No. of codons with mutations in ⩾8 tumours. For example, KRAS has four codons with a mutation in at least eight tumours (12, 13, 61 and 146). ¶HRun=MSS-associated/total homopolymer run mutations: genomic loci where a single-nucleotide base occurs numerous times consecutively in the human reference, as determined by GATK. APC and BRAF counts include all non-silent mutations as well. However, we include only APC-truncating and BRAF (V600E) mutations (that are considered as clearly functional) in the further analysis. See detailed sequencing data in Supplementary Data 1.
  3. *denotes genes with significantly lower non-silent mutation rates in TCGA samples.