Table 1 Counts of GWAS hits across runs in the UPE framework, \({{{{\mathcal{C}}}}}_{\ell }\) for each locus , which represents the number of runs for which the corresponding locus shows at least one association with P < PGW = 5 × 10−8 (see details in the Methods section)

From: Unsupervised ensemble-based phenotyping enhances discoverability of genes related to left-ventricular morphology

Chromosome

Region

Candidate gene

Count

Minimum P value

Lead variant

NEA/EA

EAF (%)

\(\bf{|} \hat{\boldsymbol{\beta} }| \pm {{{\bf{se}}}}(\hat{\boldsymbol{\beta} })({\boldsymbol{\times 1{0}^{-2}}})\)

10

120591353–122407323

BAG3

35

4.1 × 10−18

rs375034445

A/AT

21.2

5.29 ± 0.79

2

178553183–181312739

TTN

35

1.4 × 10−17

rs2042995

T/C

23.2

5.70 ± 0.77

6

117672972–118963115

PLN

35

2.0 × 10−29

rs11153730

T/C

48.6

4.29 ± 0.64

14

23018665–24905123

MYH6

34

2.7 × 10−14

rs365990

A/G

36.9

4.04 ± 0.66

12

113986709–115036602

TBX5

34

2.3 × 10−11

rs4767239

G/C

82.5

4.12 ± 0.85

12

110336719–113263518

MYL2

34

2.8 × 10−15

rs35350651

A/AC

51.4

4.36 ± 0.64

4

119933512–120392684

MYOZ2

33

2.4 × 10−13

4:120304290_GC_G

GC/G

29.0

3.96 ± 0.71

10

73508512–75422550

SYNPO2L

31

2.5 × 10−15

rs3740293

A/C

14.3

5.07 ± 0.92

3

157312028–159477890

SHOX2

30

7.0 × 10−15

rs11706187

A/G

50.1

3.23 ± 0.64

1

154770403–156336133

FDPS

29

9.7 × 10−13

rs41314549

T/C

2.81

13.7 ± 1.9

17

43056905–45876022

GOSR2

26

8.3 × 10−22

rs17608766

T/C

14.3

5.73 ± 0.90

3

99373762–100592217

FILIP1La

25

8.4 × 10−14

rs9811920

G/A

40.8

3.23 ± 0.65

16

4001196–5118345

SRL

24

9.4 × 10−12

rs889807

T/C

50.8

3.24 ± 0.64

1

21736588–23086883

KDM1A

22

2.0 × 10−12

rs35001652

G/A

37.3

2.59 ± 0.67

7

45952922–46986720

IGFBP3

22

2.0 × 10−11

rs143741275

A/AGTGTGT

42.4

2.59 ± 0.66

5

171074292–172678327

NKX2-5

21

9.0 × 10−14

rs35564079

C/CT

28.5

2.76 ± 0.72

3

13070799–14816900

TMEM43

21

2.7 × 10−11

rs900173

T/C

34.0

3.48 ± 0.68

1

144977494–148361253

GJA5

20

1.1 × 10−10

rs12046416

A/G

33.5

2.81 ± 0.68

1

235819436–237555628

ACTN2

20

2.0 × 10−12

rs12142143

T/C

53.1

4.11 ± 0.65

13

20686720–22242174

FGF9

20

8.8 × 10−13

rs10628955

G/GAA

47.7

3.95 ± 0.67

4

111256567–113870102

PITX2

19

4.4 × 10−14

rs2723294

C/T

69.5

4.65 ± 0.70

1

14891511–16897730

HSPB7

17

2.3 × 10−11

rs1763605

T/G

67.5

2.80 ± 0.68

2

146445570–147277162

ACVR2A/ZEB2

16

2.9 × 10−11

rs573709385

A/AT

44.9

1.93 ± 0.65

6

125424383–127540461

HEY2

16

2.9 × 10−11

rs11423823

C/CT

50.2

3.23 ± 0.66

2

36122006–38132712

STRN

15

9.9 × 10−16

rs2110944

T/C

52.6

3.71 ± 0.64

16

79134815–80297374

MAFa

15

5.2 × 10−11

rs558328129

A/AT

45.4

2.97 ± 0.70

8

107410754–108648177

ABRA

15

8.1 × 10−11

rs72007904

A/AACTATTC

50.0

3.87 ± 0.64

21

27271019–29125226

ADAMTS1

15

1.4 × 10−10

rs2830977

G/A

22.0

4.29 ± 0.78

1

49894177–51713726

RNF11a

13

3.1 × 10−11

rs7555411

C/T

1.42

13.8 ± 2.7

13

98938919–100574095

DOCK9a

13

1.2 × 10−10

rs34138434

C/A

29.7

4.62 ± 0.72

2

118367466–121303783

EN1a

13

6.7 × 10−14

rs162746

A/G

67.5

4.01 ± 0.68

12

27799773–29651255

CCDC91a

12

2.0 × 10−14

rs5797270

G/GT

20.3

3.35 ± 0.81

13

25784362–27284362

WASF3a

11

9.3 × 10−11

rs61944841

G/A

41.3

3.07 ± 0.67

18

19485844–20649472

GATA6

11

6.6 × 10−11

rs62094198

T/A

39.6

2.21 ± 0.66

14

19002084–21589402

NDRG2

10

2.5 × 10−11

rs12889267

A/G

16.7

3.43 ± 0.85

6

35455756–37572596

CDKN1A

10

3.2 × 10−13

rs3176326

G/A

19.8

3.47 ± 0.81

7

118351581–121045273

WNT16

10

2.4 × 10−11

rs3801387

A/G

28.1

2.79 ± 0.72

16

75977954–77523678

ADAMTS18

10

5.2 × 10−13

rs62046468

C/T

37.6

4.80 ± 0.66

17

41772087–43056905

SOST

10

5.5 × 10−11

rs17881550

G/GC

43.4

3.65 ± 0.65

11

27020461–28481593

CCDC34a

8

5.7 × 10−12

rs10835164

C/T

25.6

3.84 ± 0.74

5

120452166–122556905

PRDM6

8

6.2 × 10−11

rs463106

T/C

47.2

3.68 ± 0.65

17

67858770–69387817

KCNJ2/SOX9

7

3.6 × 10−12

rs56229089

G/C

55.7

2.89 ± 0.65

5

63968304–65911286

ADAMTS6

7

1.6 × 10−11

rs753963943

ATT/A

42.5

2.88 ± 0.66

18

8498931–11075913

NDUFV2

6

6.1 × 10−12

rs206524

T/C

70.7

3.01 ± 0.71

11

65898631–68005825

KDM2A

4

2.1 × 10−11

rs12785906

G/C

5.84

7.40 ± 1.38

1

1892607–3582736

PRDM16

3

1.7 × 10−11

rs781212641

G/GC

9.33

5.68 ± 1.12

4

7539692–8152235

AFAP1

3

1.4 × 10−11

rs28542374

G/A

63.5

3.10 ± 0.67

12

76511314–78570570

NAV3

3

2.7 × 10−12

rs7965680

T/C

55.7

2.43 ± 0.65

17

15019097–16412342

CENPVa

3

3.4 × 10−11

rs7477

A/C

49.8

3.18 ± 0.64

  1. The total number of runs was 36. The lead variant is the one for which the minimum P value occurs. P values are two-sided and derived from a linear association t-statistic (no adjustments were made for multiple comparisons). NEA and EA stand for non-effect and effect allele, respectively, whereas \(\hat{\beta }\) is the standardized effect size estimate. EAF is the frequency of EA. Note that these values correspond to different phenotypes of the ensemble, therefore they are not comparable in terms of the magnitude of the morphological change that they produce. The directions of effect can be understood with the help of Supplementary Table 5.
  2. aThe genes were annotated based purely on closest proximity to the lead variant in that region (the rest of the genes have additional previous evidence of a link to cardiac physiology as discussed in the text).
  3. Gene names that are in bold mean that the corresponding locus is only discovered with UPE, that is the remaining loci were already found by testing the handcrafted phenotypes or the shape PCs.