Supplementary Materials
Copy Number Variants and Therapeutic Response to Antidepressant Medication in Major Depressive Disorder
Supplementary Materials......
1.Materials and Methods
1.1.Samples
1.1.1.GENDEP
1.1.2.GenPod
1.1.3.GODS
1.1.4.Pfizer
1.1.5.GlaxoSmithKline
2.CNV calls per samples......
3.NEWMEDS Global CNV results......
4.Study level meta-analysis of the component studies in NEWMEDS......
4.1.Materials and Methods
4.1.1.Statistical analysis
4.2.Results
4.2.1.Whole Sample Analysis
4.2.1.1.Component studies Results
4.2.1.2.Meta Analysis Results......
4.2.2.Serotonergic Antidepressants Only Analysis
4.2.2.1.Component studies Results
4.2.2.2.Meta- Analysis Results
4.2.3.Noradrenergic Antidepressants Only Analysis
4.2.3.1.Component studies Results
4.2.3.2.Meta- Analysis Results
4.2.4.Differential Response to Treatment with either a SSRI or NRI......
4.2.4.1.Component studies Results
4.2.4.2.Meta- Analysis Results
5.Gene Set Analysis
6.Specific CNV illustrations
6.1.15q13.3......
6.2.4q28.3......
6.3.6q12......
6.4.3q26.2
6.5.20p12.1-12.2
6.6.8p23.2
6.7.2p16.3 (NXRN1)
6.8.17q25.1
6.9.18p11.32
6.10.9p23
6.11.15q13.2
6.12.3q23
7.References:
1.Materials and Methods
1.1.Samples
1.1.1.GENDEP
The Genome-Based Therapeutic Drugs for Depression (GENDEP) sample and design is described in detail elsewhere.1 Briefly, GENDEP was a twelve-week open-label part-randomized multi-centre study with two active pharmacological treatment arms with different mechanisms of antidepressant action with escitalopram primarily affecting serotonergic neurotransmission and nortriptyline primarily affecting noradrenergic neurotransmission.2,3GENDEP recruited 868 treatment-seeking adults (age 19-72 year; men n=321; women n=547) of white European origin diagnosed with ICD-10/DSM-IV unipolar major depression of at least moderate severity established in the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) interview4 with severity of depression was assessed weekly by three established rating scales.5 Personal or family history of bipolar disorder or schizophrenia and active substance dependence constituted exclusion criteria. Eligible participants were treated with either escitalopram or nortriptyline. Patients with no contraindications were randomly allocated to flexible-dosage nortriptyline (50–150 mg daily) or escitalopram (10–30 mg daily) for 12 weeks. Patients with contraindications for one drug were offered the other. Individuals were excluded from the analysis if they had no post baseline information. The study was approved by ethics boards in all participating centres. All participants provided a written consent after the procedures were explained. GENDEP is registered at EudraCT (No.2004-001723-38, and ISRCTN (No. 03693000,
1.1.2.GenPod
Detailed information for the GENetic and clinical Predictors Of treatment response in Depression (GenPod) study can be found elsewhere.6 GenPod was a multi-centre randomized clinical trial of 601 treatment seeking adults (age 18-74 years; men n=161; women n=347) recruited in primary care who had an ICD-10 diagnosis of major depression of at least moderate severity as assessed by the Clinical Interview Schedule-Revised (CIS-R)7 and the Beck Depression Inventory (BDI).8 Individuals were randomly allocated to either reboxetine (4mg twice daily) or citalopram (20mg), which represent different mechanisms of antidepressant action with citalopram primarily affecting serotonergic neurotransmission and reboxetine primarily affecting noradrenergic neurotransmission. As ethnicity is a major confounder in genetic studies due to the introduction of LD and haplotypic difference across ethnic backgrounds, only individuals with a white European ancestry were chosen for the whole genome analysis (n=512). Exclusion criteria included if individuals had psychosis, bipolar disorder or major substance or alcohol abuse, or if they had medical contraindications. Individuals were further excluded from the analysis if they had no post baseline information. All participants provided written consent after the study and procedure were explained. Ethical approval was obtained from the South West Ethics Committee (MREC 02/6/076) as well as research governance approval from Bristol, Manchester and Newcastle Primary Care NHS Trusts. GenPod is registered atEudraCT (No. 2004-001434-16) and ISRCTN (No. 31345163).
1.1.3.GODS
The Geneva Outpatient Depression Study (GODS) has been described in detail elsewhere.9, 10 Briefly, GODS is a partly randomized study, which examined the efficacy of four antidepressants (paroxetine, clomipramine, venlafaxine and nefazodone) based on a seven-step algorithm in a cohort of 131 treatment seeking adultsof white European ancestry (age 18-65 years; men n=53; women n=78) with severe MDD.11 Exclusion criteria included pregnancy, schizophrenia or schizoaffective disorder, dependence on alcohol or other substances and treatment with mood-stabilisers or antipsychotics. The present investigation includes data from the first three steps that included treatment with paroxetine (a SRI), initiated at 20mg daily and increased to 30mg and 40mg daily if remission was not achieved. Patients were discharged only if complete remission was obtained as defined by a MADRS score of 8 or less. Of the 131 GODS participants, 82 had available blood DNA samples. The study protocol was approved by the ethics committee of the Geneva University Department of Psychiatry and written informed consent was obtained from all subjects.
1.1.4.Pfizer
Pfizer contributed a total of 345 patients from eight MDD clinical trials. Study designs were variable and primarily conducted as double-blind, placebo-controlled, 6 to 8 weeks studies with sertraline, fluoxetine or paroxetine as active comparators in addition to the investigational compound. Only subjects from the SRI comparator arms were sent for whole genome genotyping and included in the current study. A diagnosis and inclusion criterion for MDD was a Hamilton Depression Rating Scale (HAM-D) total score of 22 or higher at screening. Exclusion criteria included DSM-IV diagnosis of psychotic features, bipolar I or II, or major risk for suicide. All study protocols received institutional review board (IRB) approval and informed consent was obtained from participating subjects prior to sample collection.
1.1.5.GlaxoSmithKline
The samples from GlaxoSmithKline (GSK) derived from two randomized, double-blind, placebo-controlled comparisons of the antidepressant efficacy and the effects on sexual functioning of Bupropion XL and escitalopram in outpatients with moderate to severe depression.12 Escitalopram was administered at doses of 10mg/day for the first 4 weeks and either 10mg/day, or if clinically indicated, 20mg/day from Week 5 through Week 8. Included subjects had primary diagnosis of MDD with duration at recruitment lasting 12 weeks but no greater than 2 years, and having failed to respond to two adequate trials of antidepressants in the previous 2 years. The primary outcomes for depression was the change from baseline in HAMD-17 total score, whilst the secondary outcomes were percent of subjects in remission and percent of responder (HAMD-17), plus CGI-I and CGI-S. Out of the 210 patients in the escitalopram treatment arms who completed the study, 137 treatment seeking adults(age 18-64 years; 45:55 male:female ratio) were selected based on availability of consented DNA blood sample. All patients provided written informed consent prior to any study activity and the protocol for each of the studies was approved by international review boards.
2.CNV calls per samples
Tables 1 to 3 present the summary statistics for the NEWMEDS CNV analysis with the frequency of samples with CNVs stratified by type for the whole of NEWMEDS and each contributing sample. The average size of the CNV burden is also presented for the entire NEWMEDS and each contributing sample stratified by CNV type.
Table1: Summary statistics for CNV calls made with NEWMEDS and within each individuals sample for ALL CNVs.
ALL CNVs / NEWMEDSn=1565 / GENDEP
n=694 / GenPod
n=426 / GODS
n=71 / Pfizer
n=243 / GSK
n=131
Number of Samples with a CNV / 1191
(76.10%) / 521
(75.07%) / 344
(80.75%) / 50
(70.04%) / 178
(73.25%) / 98
(74.81%)
Number of Samples with a deletion CNV / 606
(38.72%) / 271
(39.05%) / 166
(38.97%) / 26
(36.62%) / 101
(41.56%) / 42
(32.06%)
Number of Samples with a duplication CNV / 951
(60.77%) / 407
(58.65%) / 285
(66.90%) / 34
(47.89%) / 142
(58.44%) / 83
(63.36%)
Number of Samples with a rare CNV / 812
(51.88%) / 346
(49.86%) / 244
(57.28%) / 39
(54.93%) / 123
(50.62%) / 60
(45.80%)
Number of Samples with arare deletion CNV / 459
(29.33%) / 208
(29.97%) / 125
(29.34%) / 21
(29.58%) / 76
(31.28%) / 29
(22.14%)
Number of Samples with a rare duplication CNV / 573
(36.61%) / 226
(32.56%) / 185
(43.43%) / 26
(36.62%) / 86
(35.39%) / 50
(38.17%)
Number of Samples with a common CNV / 823
(52.59%) / 353
(50.86%) / 238
(55.87%) / 32
(45.07%) / 127
(52.26%) / 73
(55.73%)
Number of Samples with a common deletion CNV / 217
(13.87%) / 86
(12.39%) / 65
(15.26%) / 8
(11.27%) / 41
(16.87%) / 17
(13.98%)
Number of Samples with a common duplication CNV / 619
(39.55%) / 273
(39.34%) / 177
(41.55%) / 23
(32.39%) / 90
(37.04%) / 56
(42.75%)
Average Burden of CNVs (SD) / 371.55
(524.79) / 349.73
(587.99) / 412.38
(479.78) / 273.90
(288.11) / 396.46
(506.31) / 361.04
(425.88)
Average Burden of deletion CNVs (SD) / 123.18
(370.69) / 131.01
(492.32) / 121.76
(240.68) / 86.64
(130.56) / 122.96
(240.55) / 106.56
(231.33)
Average Burden of duplication CNVs (SD) / 248.36
(373.77) / 218.72
(328.90) / 290.62
(414.92) / 187.26
(275.20) / 273.50
(446.82) / 254.48
(339.13)
Average Burden of rare CNVs (SD) / 236.67
(492.77) / 217.81
(562.60) / 270.07
(440.42) / 168.74
(213.64) / 259.02
(471.33) / 223.33
(385.16)
Average Burden of rare deletion CNVs (SD) / 94.34
(362.85) / 106.06
(489.82) / 90.21
(220.74) / 63.29
(112.85) / 85.66
(225.01) / 78.58
(206.96)
Average Burden of rare duplication CNVs (SD) / 161.59
(347.40) / 128.29
(299.37) / 200.46
(392.60) / 123.49
(221.57) / 195.96
(420.31) / 168.55
(317.99)
Average Burden of common CNVs (SD) / 136.43
(16.78) / 133.53
(164.36) / 144.16
(162.43) / 106.96
(135.92) / 138.53
(166.42) / 138.71
(173.87)
Average Burden of common deletion CNVs (SD) / 30.40
(82.73) / 26.57
(77.42) / 33.40
(85.04) / 25.16
(78.24) / 38.39
(93.59) / 28.97
(83.02)
Average Burden of common duplication CNVs (SD) / 87.45
(123.48) / 90.80
(129.47) / 90.80
(123.14) / 67.68
(106.70) / 77.54
(112.34) / 87.86
(119.89)
Table2: Summary statistics for CNV calls made with NEWMEDS and within each individuals sample for CNVs encompassing regions annotated to harbour genes (Genic CNVs).
GENIC CNVs / NEWMEDSn=1565 / GENDEP
n=694 / GenPod
n=426 / GODS
n=71 / Pfizer
n=243 / GSK
n=131
Number of Samples with a CNV / 978
(62.50%) / 415
(59.80%) / 291
(68.31%) / 42
(59.15%) / 152
(62.55%) / 78
(59.54%)
Number of Samples with a deletion CNV / 347
(22.17%) / 144
(20.75%) / 100
(23.47%) / 17
(23.94%) / 60
(24.69%) / 26
(19.85%)
Number of Samples with a duplication CNV / 813
(51.85%) / 346
(49.86%) / 246
(57.75%) / 30
(42.25%) / 123
(50.62%) / 68
(51.91%)
Number of Samples with a rare CNV / 591
(37.76%) / 236
(34.01%) / 188
(44.13%) / 30
(42.25%) / 94
(38.68%) / 43
(32.82%)
Number of Samples with arare deletion CNV / 263
(16.81%) / 110
(15.85%) / 77
(18.08%) / 13
(18.31%) / 43
(17.70%) / 20
(15.27%)
Number of Samples with a rare duplication CNV / 460
(29.39%) / 178
(25.65%) / 153
(35.92%) / 21
(29.58%) / 72
(29.63%) / 36
(27.48%)
Number of Samples with a common CNV / 659
(42.11%) / 282
(40.63%) / 186
(43.66%) / 28
(39.44%) / 107
(44.03%) / 56
(42.45%)
Number of Samples with a common deletion CNV / 98
(6.26%) / 36
(5.19%) / 29
(6.81%) / 4
(5.63%) / 22
(9.05%) / 7
(5.34%)
Number of Samples with a common duplication CNV / 528
(33.74%) / 235
(33.86%) / 147
(34.51%) / 21
(29.58%) / 78
(32.10%) / 47
(35.88%)
Average Burden of CNVs (SD) / 281.33
(491.67) / 262.12
(564.66) / 318.65
(456.22) / 212.90
(258.35) / 299.27
(432.46) / 265.61
(366.30)
Average Burden of deletion CNVs (SD) / 78.01
(350.77) / 80.30
(471.50) / 80.50
(223.03) / 59.18
(117.48) / 77.85
(219.51) / 68.29
(187.23)
Average Burden of duplication CNVs (SD) / 203.33
(346.90) / 181.82
(310.99) / 238.15
(392.73) / 153.72
(246.45) / 221.42
(394.16) / 197.32
(312.44)
Average Burden of rare CNVs (SD) / 175.60
(460.26) / 157.61
(538.38) / 210.33
(417.16) / 123.27
(199.02) / 190.19
(397.62) / 159.24
(333.34)
Average Burden of rare deletion CNVs (SD) / 61.85
(344.79) / 66.94
(469.24) / 63.36
(207.40) / 45.29
(106.89) / 54.24
(206.57) / 53.13
(172.44)
Average Burden of rare duplication CNVs (SD) / 129.68
(320.63) / 140.28
(283.64) / 163.89
(370.12) / 93.93
(196.07) / 154.92
(363.78) / 124.97
(289.59)
Average Burden of common CNVs (SD) / 105.74
(146.19) / 104.51
(148.79) / 108.32
(144.79) / 89.62
(121.87) / 109.08
(143.24) / 106.37
(155.49)
Average Burden of common deletion CNVs (SD) / 16.16
(65.41) / 13.36
(60.41) / 17.14
(65.46) / 13.90
(60.44) / 23.61
(77.95) / 15.16
(67.60)
Average Burden of common duplication CNVs (SD) / 74.32
(115.81) / 77.80
(120.91) / 74.90
(114.99) / 63.69
(106.50) / 66.50
(105.39) / 74.28
(114.76)
Table3: Summary statistics for CNV calls made with NEWMEDS and within each individuals sample for CNVs annotated to encompass exonic regions of genes (Exonic CNVs).
EXONIC CNVs / NEWMEDSn=1565 / GENDEP
n=694 / GenPod
n=426 / GODS
n=71 / Pfizer
n=243 / GSK
n=131
Number of Samples with a CNV / 967
(61.79%) / 409
(58.93%) / 289
(67.84%) / 42
(59.15%) / 151
(62.14%) / 76
(58.02%)
Number of Samples with a deletion CNV / 326
(20.83%) / 133
(19.16%) / 96
(22.54%) / 16
(22.54%) / 58
(23.87%) / 23
(17.56%)
Number of Samples with a duplication CNV / 813
(51.95%) / 346
(49.86%) / 246
(57.75%) / 30
(42.25%) / 123
(50.62%) / 68
(51.91%)
Number of Samples with a rare CNV / 568
(36.29%) / 225
(32.42%) / 184
(43.19%) / 28
(39.44%) / 90
(37.04%) / 41
(31.30%)
Number of Samples with arare deletion CNV / 238
(15.21%) / 98
(14.12%) / 72
(16.90%) / 12
(16.90%) / 39
(16.05%) / 17
(12.98%)
Number of Samples with a rare duplication CNV / 459
(29.33%) / 177
(25.50%) / 153
(35.92%) / 21
(29.58%) / 72
(29.63%) / 36
(27.48%)
Number of Samples with a common CNV / 659
(42.11%) / 282
(40.63%) / 186
(43.66%) / 28
(39.44%) / 107
(44.03%) / 56
(42.75%)
Number of Samples with a common deletion CNV / 98
(6.26%) / 36
(5.19%) / 29
(6.81%) / 4
(5.63%) / 22
(9.05%) / 7
(5.34%)
Number of Samples with a common duplication CNV / 528
(33.74% / 235
(33.86%) / 147
(34.51%) / 21
(29.58%) / 78
(32.10%) / 47
(35.88%)
Average Burden of CNVs (SD) / 277.86
(490.53) / 258.69
(564.26) / 315.37
(453.08) / 207.50
(255.58) / 295.97
(431.68) / 261.97
(366.45)
Average Burden of deletion CNVs (SD) / 74.99
(349.68) / 77.10
(470.58) / 77.88
(221.36) / 57.48
(117.45) / 74.55
(217.13) / 64.65
(186.87)
Average Burden of duplication CNVs (SD) / 202.88
(346.76) / 181.59
(310.95) / 237.49
(392.68) / 150.02
(242.51) / 221.42
(394.16) / 197.32
(312.44)
Average Burden of rare CNVs (SD) / 172.12
(459.57) / 154.18
(538.44) / 207.05
(414.19) / 177.87
(199.34) / 186.90
(397.30) / 155.60
(33.69)
Average Burden of rare deletion CNVs (SD) / 58.83
(343.96) / 63.74
(468.44) / 60.74
(206.26) / 43.58
(106.64) / 50.94
(205.22) / 49.49
(171.73)
Average Burden of rare duplication CNVs (SD) / 129.23
(320.42) / 104.16
(283.66) / 163.23
(369.93) / 90.24
(189.91) / 154.92
(363.78) / 124.97
(289.59)
Average Burden of common CNVs (SD) / 105.74
(146.19) / 104.51
(148.79) / 108.32
(144.79) / 89.62
(121.87) / 109.08
(143.24) / 106.37
(155.49)
Average Burden of common deletion CNVs (SD) / 16.16
(65.41) / 13.36
(60.41) / 17.14
(65.46) / 13.90
(60.44) / 23.61
(77.95) / 15.16
(67.60)
Average Burden of common duplication CNVs (SD) / 74.32
(115.81) / 77.80
(120.91) / 74.90
(114.99) / 63.69
(106.50) / 66.50
(105.39) / 74.28
(114.76)
3.NEWMEDS Global CNV results
Table4: NEWMEDS CNV results for ALL CNVs. Regression coefficient is standardized and can be interpreted as a measure of effect size. Positive values of regression coefficient mean that carriers of the CNV had better treatment outcome. Negative values of regression coefficient mean that carriers had worse outcomes.
All CNVs / Whole Sample Analysis(n=1,565) / Serotonergic Analysis
(n=1,046) / Noradrenergic Analysis
(n=519) / CNV by Drug Interaction
(n=1,565)
Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value
Any CNV / -0.015
(0.058) / 0.799 / -0.0001
(0.071) / 0.999 / -0.034
(0.104) / 0.742 / -0.022
(0.126) / 0.863
Number of CNVs / -0.012
(0.021) / 0.572 / -0.022
(0.025) / 0.384 / 0.022
(0.037) / 0.557 / 0.045
(0.045) / 0.315
Burden of CNVs / 0.00002
(0.00005) / 0.739 / -0.00002
(0.00007) / 0.786 / 0.00005
(0.00006) / 0.400 / 0.00008
(0.0001) / 0.382
Any deletion CNV / -0.082
(0.051) / 0.108 / -0.108
(0.062) / 0.085 / -0.013
(0.090) / 0.884 / 0.100
(0.109) / 0.358
Number of deletion CNVs / -0.043
(0.036) / 0.232 / -0.045
(0.044) / 0.299 / -0.025
(0.062) / 0.686 / 0.022
(0.076) / 0.774
Burden of deletion CNVs / 0.000009
(0.00007) / 0.889 / -0.0002
(0.0001) / 0.286 / 0.00006
(0.00008) / 0.425 / 0.0002
(0.0002) / 0.161
Any duplication CNV / 0.038
(0.051) / 0.461 / 0.041
(0.062) / 0.514 / 0.044
(0.090) / 0.623 / 0.013
(0.109) / 0.908
Number of duplication CNVs / 0.004
(0.025) / 0.870 / -0.010
(0.031) / 0.739 / 0.042
(0.043) / 0.329 / 0.054
(0.054) / 0.313
Burden of duplication CNVs / 0.00002
(0.00008) / 0.743 / 0.00002
(0.00008) / 0.766 / 0.00004
(0.0001) / 0.759 / 0.00002
(0.0001) / 0.890
Any rare CNV / -0.012
(0.050) / 0.812 / -0.043
(0.061) / 0.482 / 0.064
(0.087) / 0.463 / 0.119
(0.106) / 0.263
Number of rare CNVs / -0.008
(0.028) / 0.789 / -0.034
(0.035) / 0.333 / 0.058
(0.049) / 0.241 / 0.096
(0.060) / 0.110
Burden of rare CNVs / 0.00003
(0.00005) / 0.577 / -0.00001
(0.00008) / 0.856 / 0.00007
(0.00007) / 0.318 / 0.00009
(0.0001) / 0.367
Any rare deletion CNV / -0.069
(0.055) / 0.209 / -0.103
(0.067) / 0.124 / 0.016
(0.096) / 0.870 / 0.121
(0.116) / 0.297
Number of rare deletion CNVs / -0.033
(0.042) / 0.436 / -0.047
(0.052) / 0.365 / 0.007
(0.071) / 0.925 / 0.053
(0.879) / 0.548
Burden of rare deletion CNVs / 0.00002
(0.00007) / 0.742 / -0.0001
(0.0002) / 0.363 / 0.00007
(0.00008) / 0.382 / 0.0002
(0.0002) / 0.207
Any rare duplication CNV / 0.020
(0.052) / 0.707 / -0.021
(0.064) / 0.741 / 0.114
(0.090) / 0.207 / 0.152
(0.110) / 0.166
Number of rare duplication CNVs / 0.015
(0.036) / 0.664 / -0.017
(0.043) / 0.699 / 0.091
(0.062) / 0.140 / 0.115
(0.076) / 0.128
Burden of rare duplication CNVs / 0.00003
(0.00007) / 0.663 / 0.00003
(0.00009) / 0.744 / 0.00005
(0.0001) / 0.686 / 0.00004
(0.0002) / 0.823
Any common CNV / -0.003
(0.050) / 0.950 / 0.031
(0.061) / 0.610 / -0.069
(0.087) / 0.429 / -0.088
(0.106) / 0.405
Number of common CNVs / -0.011
(0.032) / 0.721 / -0.004
(0.039) / 0.927 / -0.016
(0.053) / 0.771 / -0.016
(0.066) / 0.808
Burden of common CNVs / -0.00008
(0.0002) / 0.607 / -0.00004
(0.0002) / 0.824 / -0.0001
(0.0003) / 0.707 / -0.00007
(0.0003) / 0.839
Any common deletion CNV / -0.048
(0.072) / 0.505 / -0.005
(0.087) / 0.950 / -0.136
(0.128) / 0.289 / -0.124
(0.156) / 0.428
Number of common deletion CNVs / -0.035
(0.064) / 0.583 / -0.013
(0.078) / 0.870 / -0.075
(0.116) / 0.518 / -0.059
(0.140) / 0.671
Burden of common deletion CNVs / -0.0002
(0.0003) / 0.529 / -0.0002
(0.0004) / 0.657 / -0.0002
(0.0005) / 0.770 / 0.00004
(0.0006) / 0.955
Any common duplication CNV / 0.013
(0.051) / 0.799 / 0.025
(0.062) / 0.694 / -0.006
(0.088) / 0.941 / -0.025
(0.108) / 0.820
Number of common duplication CNVs / -0.008
(0.038) / 0.840 / -0.003
(0.048) / 0.952 / -0.005
(0.062) / 0.938 / -0.007
(0.079) / 0.929
Burden of common duplication CNVs / -0.00004
(0.0002) / 0.833 / 0.000002
(0.0003) / 0.995 / -0.00007
(0.0003) / 0.842 / -0.00008
(0.0004) / 0.844
Table5: NEWMEDS CNV results for CNVs encompassing regions annotated to harbour genes. Regression coefficient is standardized and can be interpreted as a measure of effect size. Positive values of regression coefficient mean that carriers of the CNV had better treatment outcome. Negative values of regression coefficient mean that carriers had worse outcomes.
GENIC CNVs / Whole Sample Analysis(n=1,565) / Serotonergic Analysis
(n=1,046) / Noradrenergic Analysis
(n=519) / CNV by Drug Interaction
(n=1,565)
Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value
Any CNV / -0.042
(0.051) / 0.415 / -0.049
(0.063) / 0.433 / -0.012
(0.091) / 0.895 / 0.049
(0.110) / 0.658
Number of CNVs / -0.008
(0.025) / 0.750 / -0.016
(0.030) / 0.606 / 0.018
(0.043) / 0.674 / 0.041
(0.052) / 0.438
Burden of CNVs / 0.00003
(0.00005) / 0.617 / 0.00001
(0.00009) / 0.898 / 0.00004
(0.00007) / 0.519 / 0.00005
(0.0001) / 0.652
Any deletion CNV / -0.082
(0.060) / 0.172 / -0.068
(0.073) / 0.354 / -0.098
(0.105) / 0.350 / -0.002
(0.128) / 0.986
Number of deletion CNVs / -0.085
(0.048) / 0.080 / -0.066
(0.060) / 0.275 / -0.108
(0.082) / 0.188 / -0.023
(0.102) / 0.821
Burden of deletion CNVs / 0.0000005
(0.00007) / 0.994 / -0.0001
(0.0002) / 0.382 / 0.00004
(0.00008) / 0.648 / 0.0002
(0.0002) / 0.272
Any duplication CNV / 0.019
(0.050) / 0.701 / -0.008
(0.061) / 0.893 / 0.091
(0.087) / 0.298 / 0.101
(0.106) / 0.339
Number of duplication CNVs / 0.019
(0.029) / 0.506 / 0.001
(0.035) / 0.972 / 0.064
(0.050) / 0.197 / 0.065
(0.061) / 0.284
Burden of duplication CNVs / 0.00005
(0.00007) / 0.482 / 0.00005
(0.00009) / 0.545 / 0.00006
(0.0001) / 0.624 / 0.00002
(0.0002) / 0.898
Any rare CNV / 0.006
(0.052) / 0.905 / 0.001
(0.063) / 0.985 / 0.031
(0.089) / 0.729 / 0.044
(0.109) / 0.689
Number of rare CNVs / 0.005
(0.034) / 0.877 / -0.009
(0.043) / 0.834 / 0.044
(0.058) / 0.440 / 0.067
(0.072) / 0.350
Burden of rare CNVs / 0.00004
(0.00005) / 0.446 / 0.00003
(0.00009) / 0.717 / 0.00005
(0.00007) / 0.452 / 0.00004
(0.0001) / 0.748
Any rare deletion CNV / -0.057
(0.067) / 0.390 / -0.033
(0.081) / 0.681 / -0.105
(0.117) / 0.371 / -0.044
(0.143) / 0.760
Number of rare deletion CNVs / -0.068
(0.055) / 0.217 / -0.040
(0.069) / 0.557 / -0.112
(0.091) / 0.222 / -0.051
(0.115) / 0.658
Burden of rare deletion CNVs / 0.00001
(0.00007) / 0.0858 / -0.0001
(0.0002) / 0.572 / 0.00004
(0.00008) / 0.649 / 0.0002
(0.0002) / 0.434
Any rare duplication CNV / 0.038
(0.055) / 0.490 / -0.011
(0.068) / 0.866 / 0.153
(0.093) / 0.103 / 0.177
(0.115) / 0.126
Number of rare duplication CNVs / 0.042
(0.041) / 0.301 / 0.009
(0.050) / 0.861 / 0.123
(0.071) / 0.081 / 0.122
(0.087) / 0.161
Burden of rare duplication CNVs / 0.00006
(0.00008) / 0.411 / 0.00007
(0.0001) / 0.489 / 0.00008
(0.001) / 0.577 / 0.00002
(0.0002) / 0.895
Any common CNV / -0.026
(0.050) / 0.601 / -0.041
(0.062) / 0.502 / 0.009
(0.088) / 0.915 / 0.062
(0.107) / 0.563
Number of common CNVs / -0.025
(0.038) / 0.510 / -0.028
(0.048) / 0.562 / -0.013
(0.063) / 0.831 / 0.014
(0.079) / 0.859
Burden of common CNVs / -0.0001
(0.0002) / 0.476 / -0.0001
(0.0002) / 0.582 / -0.00009
(0.0003) / 0.750 / 0.00003
(0.0004) / 0.928
Any common deletion CNV / -0.139
(0.103) / 0.177 / -0.149
(0.128) / 0.244 / -0.083
(0.173) / 0.631 / 0.079
(0.215) / 0.712
Number of common deletion CNVs / -0.135
(0.100) / 0.175 / -0.143
(0.122) / 0.242 / -0.083
(0173) / 0.631 / 0.072
(0.211) / 0.734
Burden of common deletion CNVs / -0.0003
(0.0004) / 0.365 / -0.0005
(0.0005) / 0.326 / 0.00003
(0.0006) / 0.966 / 0.0005
(0.0008) / 0.500
Any common duplication CNV / 0.005
(0.053) / 0.920 / -0.003
(0.064) / 0.968 / 0.018
(0.092) / 0.843 / 0.030
(0.112) / 0.787
Number of common duplication CNVs / -0.003
(0.042) / 0.945 / -0.006
(0.053) / 0.913 / 0.006
(0.069) / 0.926 / 0.011
(0.087) / 0.896
Burden of common duplication CNVs / -0.00003
(0.0002) / 0.883 / -0.00002
(0.0003) / 0.946 / -0.00003
(0.0004) / 0.931 / -0.000009
(0.0005) / 0.985
Table6: NEWMEDS CNV results for CNVs annotated to encompass exonic regions of genes. Regression coefficient is standardized and can be interpreted as a measure of effect size. Positive values of regression coefficient mean that carriers of the CNV had better treatment outcome. Negative values of regression coefficient mean that carriers had worse outcomes.
EXONIC CNVs / Whole Sample Analysis(n=1,565) / Serotonergic Analysis
(n=1,046) / Noradrenergic Analysis
(n=519) / CNV by Drug Interaction
(n=1,565)
Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / P-value
Any CNV / -0.028
(0.051) / 0.591 / -0.038
(0.062) / 0.546 / 0.008
(0.090) / 0.929 / 0.054
(0.109) / 0.623
Number of CNVs / -0.002
(0.025) / 0.943 / -0.014
(0.031) / 0.656 / 0.032
(0.043) / 0.458 / 0.053
(0.053) / 0.322
Burden of CNVs / 0.00003
(0.00005) / 0.547 / 0.00001
(0.00008) / 0.855 / 0.00005
(0.00007) / 0.466 / 0.00005
(0.0001) / 0.644
Any deletion CNV / -0.064
(0.061) / 0.300 / -0.070
(0.075) / 0.346 / -0.036
(0.108) / 0.737 / 0.057
(0.131) / 0.661
Number of deletion CNVs / -0.073
(0.051) / 0.148 / -0.067
(0.063) / 0.285 / -0.075
(0.086) / 0.387 / 0.012
(0.107) / 0.911
Burden of deletion CNVs / 0.000008
(0.00007) / 0.916 / -0.0001
(0.0002) / 0.404 / 0.00004
(0.00008) / 0.589 / 0.0002
(0.0002) / 0.274
Any duplication CNV / 0.019
(0.050) / 0.701 / -0.008
(0.061) / 0.893 / 0.091
(0.087) / 0.298 / 0.102
(0.106) / 0.339
Number of duplication CNVs / 0.021
(0.029) / 0.460 / 0.003
(0.035) / 0.930 / 0.067
(0.050) / 0.178 / 0.066
(0.061) / 0.276
Burden of duplication CNVs / 0.00005
(0.00007) / 0.457 / 0.00006
(0.00008) / 0.521 / 0.00007
(0.0001) / 0.608 / 0.00002
(0.0002) / 0.898
Any rare CNV / 0.025
(0.052) / 0.630 / 0.010
(0.064) / 0.878 / 0.068
(0.090) / 0.446 / 0.073
(0.110) / 0.508
Number of rare CNVs / 0.018
(0.035) / 0.611 / -0.004
(0.044) / 0.918 / 0.071
(0.059) / 0.228 / 0.089
(0.073) / 0.222
Burden of rare CNVs / 0.00005
(0.00005) / 0.385 / 0.00004
(0.00009) / 0.672 / 0.00006
(0.00007) / 0.401 / 0.00004
(0.0001) / 0.744
Any rare deletion CNV / -0.027
(0.069) / 0.702 / -0.030
(0.084) / 0.723 / -0.021
(0.125) / 0.869 / 0.037
(0.150) / 0.806
Number of rare deletion CNVs / -0.050
(0.058) / 0.389 / -0.039
(0.072) / 0.592 / -0.068
(0.097) / 0.481 / -0.008
(0.121) / 0.945
Burden of rare deletion CNVs / 0.00002
(0.00007) / 0.780 / -0.00009
(0.0002) / 0.603 / 0.00004
(0.00008) / 0.589 / 0.0002
(0.0002) / 0.442
Any rare duplication CNV / 0.041
(0.055) / 0.451 / -0.011
(0.068) / 0.866 / 0.163
(0.094) / 0.082 / 0.187
(0.115) / 0.105
Number of rare duplication CNVs / 0.047
(0.041) / 0.253 / 0.013
(0.051) / 0.801 / 0.129
(0.071) / 0.068 / 0.124
(0.087) / 0.154
Burden of rare duplication CNVs / 0.00007
(0.00009) / 0.386 / 0.00007
(0.0001) / 0.464 / 0.00008
(0.0001) / 0.561 / 0.00002
(0.0002) / 0.897
Any common CNV / -0.026
(0.050) / 0.601 / -0.041
(0.062) / 0.502 / 0.009
(0.088) / 0.915 / 0.062
(0.107) / 0.563
Number of common CNVs / -0.025
(0.038) / 0.510 / -0.028
(0.048) / 0.562 / -0.013
(0.063) / 0.831 / 0.014
(0.079) / 0.859
Burden of common CNVs / -0.0001
(0.0002) / 0.476 / -0.0001
(0.0002) / 0.582 / -0.00009
(0.0003) / 0.750 / 0.00003
(0.0004) / 0.928
Any common deletion CNV / -0.139
(0.103) / 0.177 / -0.149
(0.128) / 0.244 / -0.083
(0.173) / 0.631 / 0.079
(0.215) / 0.712
Number of common deletion CNVs / -0.135
(0.100) / 0.175 / -0.143
(0.122) / 0.242 / -0.083
(0.173) / 0.631 / 0.072
(0.211) / 0.734
Burden of common deletion CNVs / -0.0004
(0.0004) / 0.365 / -0.0005
(0.0005) / 0.326 / 0.00003
(0.0006) / 0.966 / 0.0005
(0.0008) / 0.500
Any common duplication CNV / 0.005
(0.053) / 0.920 / -0.003
(0.064) / 0.968 / 0.018
(0.092) / 0.842 / 0.030
(0.112) / 0.787
Number of common duplication CNVs / -0.003
(0.042) / 0.945 / -0.006
(0.053) / 0.913 / 0.006
(0.069) / 0.926 / 0.011
(0.087) / 0.896
Burden of common duplication CNVs / -0.00003
(0.0002) / 0.883 / -0.00002
(0.0003) / 0.946 / -0.00003
(0.0004) / 0.931 / -0.000009
(0.0005) / 0.985
4.Study level meta-analysis of the component studies in NEWMEDS
The main analysis of this manuscript investigated the relationship between CNVs and antidepressant response in NEWMEDS as a whole. We also investigated this relationship within each contributing study of NEWMEDS (GENDEP, GenPod, GODS, Pfizer, GlaxoSmithKline) and meta-analysed the results. The methodology and results from this analysis are presented here.
4.1.Materials and Methods
CNV calling, sample and CNV quality control all proceeded exactly as was done in the main analysis. PLINK v1.0713 was used to determine number of CNVs and total size of CNVs for each individual as was done in the main analysis. We investigated all CNVs, common CNVs (found in more than 1% of individuals) and rare CNVs (found in less than 1% of individuals) as well as the effects of harbouring any CNVs, anydeletion CNVs or any duplication CNVs. CNVs were further annotated as to whether they covering gene-coding regions (genic) or exon-coding regions (exonic) as defined by RefSeq gene annotation coordinates obtainable from the UCSC genome browser (
4.1.1.Statistical analysis
Statistical analyses were done using STATA/SE 10.14We explored the relationship between CNVs and antidepressant response within each contributing study using linear mixed models fitted with maximum restricted likelihood, as we have done previously for candidate gene analyses.1Linear mixed models are highly advantageous as they make the maximal use of all available data on treatment response by using all repeated measurement information for each individual.
For each study, fixed effects of baseline depression severity, age, sex and time (linear and quadratic), and random effects of individual and recruitment centre were included.
Results from the individuals studies within NEWMEDS were meta-analyzed using the meta command in STATA/SE 10.14 Both fixed effects and random effects models were tested.
4.2.Results
4.2.1.Whole Sample Analysis
4.2.1.1.Component studies Results
Linear mixed models were run for each contributing study of NEWMEDS using STATA/SE 1014 with fixed effects of baseline depression severity, age, sex and time (linear and quadratic), and random effects of individual and recruitment centre were included. Results from each of the individual contributing study with NEWMEDS for individuals given any antidepressant (whole sample analysis) for all, genic and exonic CNVS are presented in Tables 4,5 and 6.
Table7: Results from each component study for individuals given any antidepressant (whole sample analysis) for ALL CNVS. Regression coefficient is standardized and can be interpreted as effect size. Positive values of regression coefficient mean that carriers of more minor alleles had better treatment outcome. Negative values of regression coefficient mean that carriers of more minor alleles had worse outcomes. Results meeting the alpha level of p<0.05 are highlighted in bold and italicised.
ALL CNVs / GENDEP ALL / GenPod ALL / GODS ALL / PFZ ALL / GSK ALLtype / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value / Coefficient
(SE) / p-value
Number of CNVs / -0.283
(0.165) / 0.086 / 0.122
(0.234) / 0.602 / 0.743
(0.589) / 0.207 / -0.00134
(0.014) / 0.925 / 0.053
(0.286) / 0.854
Burden of CNVs / -0.000822
(0.000316) / 0.009 / 0.0000541
(0.000609) / 0.929 / 0.000324
(0.00228) / 0.887 / -0.00000741
(0.000036) / 0.837 / 0.000128
(0.000827) / 0.877
Number of deletion
CNVs / -0.271
(0.282) / 0.338 / 0.300
(0.396) / 0.448 / 1.122
(1.032) / 0.277 / 0.00619
(0.025) / 0.808 / 0.564
(0.498) / 0.258
Burden of deletion CNVs / -0.000908
(0.000378) / 0.016 / 0.000129
(0.00123) / 0.916 / 0.00327
(0.0051) / 0.522 / 0.00000972
(0.0000754) / 0.897 / 0.0016
(0.00149) / 0.283
Number of duplication
CNVs / -0.284
(0.202) / 0.159 / -0.030
(0.278) / 0.914 / 0.455
(0.653) / 0.486 / -0.0052
(0.018) / 0.772 / -0.233
(0.380) / 0.540
Burden of duplication
CNVs / -0.000607
(0.000568) / 0.285 / -0.0000278
(0.000706) / 0.969 / -0.000362
(0.00024) / 0.880 / -0.0000126
(0.0000412) / 0.760 / -0.000575
(0.00104) / 0.582
Number of rare CNVs / -0.380
(0.226) / 0.092 / 0.107
(0.310) / 0.730 / 0.867
(0.796) / 0.276 / -0.0053
(0.020) / 0.786 / 0.545
(0.418) / 0.193
Burden of rare CNVs / -0.000821
(0.00033) / 0.013 / -0.0000149
(0.000665) / 0.982 / -0.000308
(0.00285) / 0.914 / -0.0000127
(0.0000388) / 0.743 / 0.000522
(0.000909) / 0.565
Number of rare deletion
CNVs / -0.478
(0.315) / 0.127 / 0.044
(0.470) / 0.925 / 0.976
(1.239) / 0.431 / -0.0034
(0.033) / 0.917 / 0.484
(0.625) / 0.438
Burden of rare deletion
CNVs / -0.000945
(0.00038) / 0.013 / -0.000266
(0.00133) / 0.842 / 0.00148
(0.00592) / 0.802 / -0.00000433
(0.000081) / 0.957 / 0.00138
(0.00166) / 0.409
Number of rare duplication
CNVs / -0.437
(0.303) / 0.149 / 0.158
(0.374) / 0.672 / 0.429
(0.942) / 0.649 / 0.00107
(0.023) / 0.963 / 0.247
(0.542) / 0.649
Burden of rare duplication
CNVs / -0.000519
(0.000623) / 0.404 / 0.0000987
(0.000749) / 0.895 / -0.00151
(0.00299) / 0.614 / -0.0000075
(0.0000441) / 0.865 / -0.000144
(0.00111) / 0.897
Number of common CNVs / -0.220
(0.236) / 0.352 / 0.152
(0.369) / 0.681 / 0.753
(0.918) / 0.412 / 0.00385
(0.023) / 0.867 / -0.475
(0.416) / 0.253
Burden of common CNVs / -0.00109
(0.00115) / 0.342 / 0.000973
(0.0018) / 0.588 / 0.00286
(0.00492) / 0.561 / 0.0000326
(0.000109) / 0.764 / -0.00195
(0.00201) / 0.330
Number of common
deletion CNVs / 0.118
(0.521) / 0.821 / 1.263
(0.749) / 0.092 / 1.802
(1.808) / 0.319 / 0.022
(0.042) / 0.596 / 0.588
(0.896) / 0.511
Burden of common
deletion CNVs / 0.0004
(0.00245) / 0.870 / 0.00538
(0.00349) / 0.123 / 0.00805
(0.0089) / 0.366 / 0.0000899
(0.000195) / 0.644 / 0.00324
(0.00423) / 0.443
Number of common
duplication CNVs / -0.173
(0.276) / 0.530 / -0.185
(0.432) / 0.668 / 0.865
(1.187) / 0.466 / -0.017
(0.030) / 0.583 / -0.816
(0.538) / 0.129
Burden of common
duplication CNVs / -0.00119
(0.00146) / 0.414 / -0.000674
(0.00235) / 0.775 / 0.00483
(0.00651) / 0.458 / -0.0000931
(0.000162) / 0.565 / -0.00411
(0.00287) / 0.153
Table8: Results of component studies study for individuals given any antidepressant (whole sample analysis) for CNVs encompassing regions annotated to harbour genes (Genic CNVs). Regression coefficient is standardized and can be interpreted as a measure of effect size. Positive values of regression coefficient mean that carriers of more minor alleles had better treatment outcome. Negative values of regression coefficient mean that carriers of more minor alleles had worse outcomes. Results meeting the alpha level of p<0.05 are highlighted in bold and italicised.