Table 2.
All Mendelian randomization (MR) studies included for qualitative synthesis, with their identifying information, description of the data samples used for exposure and outcome variable(s), ancestry of those samples, the independence of the include SNPs, whether or not proxies were used, and whether or not a correction for multiple testing was applied
| ID | Author year | GWAS sample exposure variable(s) | Ancestry exposure sample | GWAS sample outcome variable(s) | Ancestry outcome sample | Independence of the SNPs (LD threshold or otherwise) | Proxies used, and if so, LD | Correction multiple testing |
|---|---|---|---|---|---|---|---|---|
| 1 | Zhou et al. (2019a) | Okbay et al. (2016), N = 293 723 | European | Thorgeirsson et al. (2010) (Tobacco and Genetics (TAG) consortium), smoking initiation effective-N = 72 710, smoking cessation effective-N = 41 278 | European | Independent SNPs as reported in exposure GWAS were selected | Yes, LD r2 > 0.8 | None |
| 2 | Zeng et al. (2019) | Okbay et al. (2016), N = 293 723; Lee et al. (2018), N = 1 131 881 | European | Thorgeirsson et al. (2010) (TAG consortium) effective-N = 72 710 | European | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 3 | Gage et al. (2018) | Okbay et al. (2016), N = 305 072 (Discovery and replication sample, without 23andme) | European | Thorgeirsson et al. (2010) (TAG consortium), smoking initiation effective-N = 72 710, cigarettes per day – N = 38 181, smoking cessation effective-N = 41 278; Ware et al. (2016) – N = 4548 | European | Independent SNPs as reported in exposure GWAS were selected | Yes, LD r2 > 0.9 | None |
| 4 | Tillmann et al. (2017) | Okbay et al. (2016), N = 349 306 | European | Thorgeirsson et al., 2010 (TAG consortium) effective-N = 72 710 | European | LD r2 < 0.1 | Yes, LD r2 > 0.8 | None |
| 5 | Carter et al. (2019) | Individual-level MR: UKB (UK Biobank), N = 318 147; Summary-level MR: Lee et al. (2018), N = 1 131 881 | European | Individual-level MR: UKB, N = 318 147; Summary-level MR: Wootton et al. (2019), N = 462 690 | European | LD r2 < 0.1 education; LD r2 < 0.001 smoking | No | None |
| 6 | Sanderson et al. (2019) | Individual-level MR: UKB, N = 120 050; Summary-level MR: Lee et al. (2018), N = 1 131 881 | European | Individual-level MR: UKB, N = 120 050; Summary-level MR: Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710; smoking cessation effective-N = 41 278 | European | LD r2 < 0.001 | No | None |
| 7 | Gage et al. (2020) | Liu et al. (2019a (GSCAN consortium) without UK Biobank (since it is not explicitly mentioned that these were excluded, it is assumed that 23andme data were included), smoking initiation, N = 848 460; Wootton et al. (2019), N = 462 690 | European | Okbay et al. (2016), N = 293 723; Cognitive functioning UKB (Neale lab GWAS: http://www.nealelab.is/uk-biobank), N = 117 131 | European | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 8a | Fu et al. (2019) | The Health and Retirement Cohort, N = 11 246 | European | The Health and Retirement Cohort, N = 11 246 | European | – | – | None |
| 9 | North et al. (2015) | Healthy Aging across the Life Course (HALCyon) consortium, N = 22 329 | European | HALCyon consortium, N = 22 329 | European | n.a. (1 SNP) | No | Bonferroni |
| 10a | Adams (2019) | Wootton et al., 2019, N = 462 690; Cognitive functioning UKB, N = 149 051; Okbay et al. (2016), N = 170 911 | European | Wootton et al. (2019), N = 462 690; Cognitive functioning UKB, N = 149 051; Okbay et al. (2016), N = 170 911 | European | LD r2 < 0.01 | No | False Discovery Rate (FDR) |
| 11 | Østergaard et al. (2015) | Thorgeirsson et al. (2010) (TAG consortium), smoking initiation effective-N = 72 710, cigarettes per day – N = 38 181 | European | Lambert et al. (2013) (International Genomics of Alzheimer's disease's Project (IGAP)) effective-N = 46 668 | European | LD r2 < 0.01 | Yes, LD r2 > 0.8 | Bonferroni |
| 12 | Gibson et al. (2019) | Summary-level MR: Jones et al. (2016) – N = 128 266; Hammerschlag et al. (2017), effective-N = 92 415; Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710, cigarettes per day – N = 38 181, smoking cessation effective-N = 41 278; Individual-level MR: UKB – N = 335 708 participants (computed as the biggest sample from Table 1: 184 184 + 118 181 + 33 343) | European | Summary-level MR: Jones et al. (2016), N = 128 266; Hammerschlag et al. (2017), effective-N = 92 415; Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710, cigarettes per day – N = 38 181, smoking cessation effective-N = 41 278; Individual-level MR: UKB – N = 335 708 participants (computed as the biggest sample from Table 1: 184 184 + 118 181 + 33 343) | European | LD r2 < 0.001 | Yes, LD r2 ⩾ 0.8 | None |
| 13 | Millard et al. (2019) | UKB, N = 182 961, never smokers; N = 150 831, ever smokers | European | UKB, N = 182 961, never smokers; N = 150 831, ever smokers | European | n.a. (1 SNP) | No | Bonferonni & FDR (note that this is a PHEWAS) |
| 14 | Jansen et al. (2019) | Jansen et al. (2019), N = 1.3 million; Thorgeirsson et al. (2010) (TAG consortium), N = 38 181 | European | Jansen et al. (2019), N = 1.3 million; Thorgeirsson et al. (2010) (TAG consortium), N = 38 181 | European | LD r2 < 0.1 | No | Bonferroni |
| 15 | Lane et al. (2019) | Frequent insomnia symptoms UKB effective-N = 235 787; Any insomnia symptoms UKB effective-N = 329 839 | European | Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710, cigarettes per day – N = 38 181, smoking cessation effective-N = 41 278, age at initiation – N = 24 114 | European | LD r2 > 0.8 | No | Bonferroni |
| 16 | Bjorngaard et al. (2013) | The Trøndelag Health Study (HUNT) Cohort, N = 53 601 | European | HUNT cohort, N = 53 601 | European | n.a. (1 SNP) | No | None |
| 17 | Lewis et al. (2011) | The Avon Longitudinal Study of Parents and Children (ALSPAC) Cohort, N = 6.294 | European | ALSPAC cohort, N = 6.294 | European | n.a. (1 SNP) | No | None |
| 18 | Taylor et al. (2014a) | The Causal Analysis Research in Tobacco and Alcohol (CARTA) consortium, N = 127 632 (58 176 never smokers, 37 428 former smokers, 32 028 current smokers) | European | CARTA consortium, N = 127 632 (58 176 never smokers, 37 428 former smokers, 32 028 current smokers) | European | n.a. (1 SNP) | Yes, rs16 969968 or proxy rs1051730 | None |
| 19 | Skov-Ettrup et al. (2017) | The Copenhagen General Population study (CGP), N = 90 108 | European | The Copenhagen General Population study (CGP), N = 90 108 | European | n.a. (1 SNP) | No | None |
| 20 | Wootton et al. (2019) | Wootton et al. (2019) UKB, N = 462 690; Liu et al. (2019a) (GSCAN consortium) smoking initiation, N = 1.2 million; Schizophrenia working group PGC, 2014 effective-N = 111 486; Wray et al. (2018) effective-N = 374 559 | Predominantly European (small Asian cohorts in schiz GWAS) | Wootton et al. (2019) UKB, N = 462 690; Liu et al. (2019a) (GSCAN consortium); smoking initiation, N = 599 289; Schizophrenia working group PGC, 2014 effective-N = 111 486; Wray et al. (2018) effective-N = 141 380 | Predominantly European (small Asian cohorts in schiz GWAS) | LD r2 < 0.001 | Yes, LD r2 ⩾ 0.8 | None |
| 21 | Vermeulen et al. (2019) | Liu et al. (2019a) (GSCAN consortium) smoking initiation, N = 1.2 million; Wootton et al. (2019), N = 462 690; Stahl et al. (2019) effective-N = 49 367 | European | Liu et al. (2019a) (GSCAN consortium) smoking initiation, N = 1.2 million, cigarettes per day, N = 263 954; smoking cessation, N = 312 821; Wootton et al. (2019), N = 462 690; Stahl et al. (2019) effective-N = 49 367 | European | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 22a | Barkhuizen et al. (2020) | Liu et al. (2019a) (GSCAN consortium) without 23andMe, N = 632 802; Pain et al. (2018), N = 6297–10 098; Ortega-Alonso et al. (2017), N = 3967–4057; UKB (Neale lab: http://www.nealelab.is/uk-biobank), N = 157 397; Schizophrenia working group of PGC, 2014 effective-N = 111 487; Wray et al. (2018) excluding 23andMe effective-N = 156 582; Stahl et al. (2019) effective-N = 49 367 | Predominantly European (small Asian cohorts in schiz GWAS) | Liu et al. (2019a) (GSCAN consortium) without 23andMe, N = 632 802; Pain et al. (2018), N = 6297–10 098; Ortega-Alonso et al. (2017), N = 3967–4057; UKB (Neale lab: http://www.nealelab.is/uk-biobank), N = 157 397; Schizophrenia working group of PGC, 2014 effective-N = 111 487; Wray et al. (2018) excluding 23andMe effective-N = 156 582; Stahl et al. (2019) effective-N = 49 367 | Predominantly European (small Asian cohorts in schiz GWAS) | Independent SNPs as reported in exposure GWAS were selected or LD r2 < 0.05 | No | None for MR (the genetic correlations are corrected for multiple testing) |
| 23 | Wium-Andersen et al. (2015a) | Copenhagen General Population Study (CGPS) and Copenhagen City Heart Study (CCHS) cohorts, N = 63 296 (23 282 never smokers and 40 014 ever smokers) | European | Copenhagen General Population Study (CGPS) and Copenhagen City Heart Study (CCHS) cohorts, N = 63 296 (23 282 never smokers and 40 014 ever smokers) | European | n.a. (1 SNP) | No | None |
| 24 | Byrne et al. (2019) | UKB, N = 32 510 | European | Schizophrenia working group of PGC, 2014 effective-N = 99 863 (40 675 cases and 64 643 controls) (note that the PGC schizophrenia working group is referenced but the sample size does not match that in the 2014 PGC publication) | European | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 25 | Gage et al. (2017b) | Schizophrenia working group of PGC, 2014 effective-N = 111 487; Thorgeirsson et al. (2010) (TAG consortium) effective-N = 72 710 | Predominantly European (small Asian cohorts in schiz GWAS) | Schizophrenia working group of PGC, 2014 effective-N = 111 487; Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710, cigarettes per day – N = 38 181, smoking cessation effective-N = 41 278 | Predominantly European (small Asian cohorts in schiz GWAS) | LD r2 < 0.9 | Yes, LD r2 > 0.9 | None |
| 26 | Fluharty et al. (2018) | Demontis et al. (2019) effective-N = 51 205 (note that the reported sample size implies that a small Asian cohort was included); Pappa et al. (2016) (Early Life Epidemiology consortium (EAGLE)) – N = 18 988 | Predominantly European (small Asian cohort in attention-deficit hyperactivity disorder GWAS) | Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710; age at initiation – N = 24 114 | European | Independent SNPs as reported in exposure GWAS were selected | Yes, LD r2 ≥ 0.9 | None |
| 27 | Sallis et al. (2019) | Individual-level MR: UKB, N = 273 516; summary-level MR: Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710, cigarettes per day – N = 38 181, smoking cessation effective-N = 41 278; Okbay et al. (2016) – N = 170 911; Lo et al. (2017) – N = 122 886 | European | Individual-level MR: UKB – N = 273 516; summary-level MR: Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710, cigarettes per day – N = 38 181, smoking cessation effective-N = 41 278; Okbay et al. (2016) – N = 170 911; Lo et al. (2017) – N = 122 886 | European | Independent SNPs as reported in exposure GWAS were selected | Yes LD r2 ≥ 0.8 | None |
| 28a | Leppert et al. (2019) | Demontis et al. (2019) effective-N = 49 017 | European | Wootton et al. (2019), N = 462 690 | European | LD r2 < 0.001 | Yes, LD r2 ≥ 0.9 | None |
| 29 | Harrison et al. (2020a) | Individual-level MR: UKB, N = 463 033/2 (split-sample analyses); Summary-level MR: Liu et al. (2019a) (GSCAN consortium) without UKB and 23andMe – N = 249 171; Wootton et al. (2019) – N = 463 033 | European | Individual-level MR: UKB – N = 463 033/2 (split-sample analyses); Summary-level MR: UKB – N = effective-N = 9661 (2433 cases and 334 766 controls) | European | Independent SNPs as reported in exposure GWAS were selected | Yes, LD r2 ≥ 0.8 | None |
| 30 | Rosoff et al. (2019) | Okbay et al. (2016) – N = 293 723; Elsworth et al. (2017) alcohol intake frequency – N = 462 346, weekly intake – N = 326 801; Karlsson Linnér et al. (2019) – N = 414 343; Walters et al. (2018) – N = 28 657; Sanchez-Roige et al. (2019) – N = 121 604 | European | Okbay et al. (2016) – N = 293 723; Elsworth et al. (2017) alcohol intake frequency – N = 462 346, weekly intake – N = 326 801; Karlsson Linnér et al. (2019) – N = 414 343; Walters et al. (2018) – N = 28 657; Sanchez-Roige et al. (2019) – N = 121 604 | European | LD r2 = 0.001 | It is mentioned a proxy was used for one SNP but not the LD | Bonferroni |
| 31 | Zhou et al. (2019b) | Lee et al. (2018) – N = 1 131 881 | European | UKB, N = 334 507 | European | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 32 | Kumari et al. (2014) | English Longitudinal Study of Aging (ELSA) + Whitehall II study + Health, Alcohol and Psychosocial factors in Easter Europe Study (HAPIEE) combined, N = 34 452 | European | ELSA + Whitehall II study + HAPIEE combined, N = 34 452 | European | n.a. (1 SNP) | No | None |
| 33 | Almeida et al. (2014a) | The Health in Men Study (HIMS) Cohort, N = 3542 | Predominantly European | HIMS Cohort, N = 3542 | Predominantly European | n.a. (1 SNP) | No | None |
| 34 | Ritchie et al. (2014) | The Lothian Birth Cohort 1936, N = 777 | European | The Lothian Birth Cohort 1936, N = 777 | European | Four SNPs as previously reported in candidate-gene literature | No | None |
| 35 | Au Yeung et al. (2012) | The Guangzhou Biobank Cohort Study (GBCS), N = 4707 | Chinese | GBCS, N = 4707 | Chinese | n.a. (1 SNP) | No | None |
| 36 | Mahedy et al. (2020) | Liu et al. (2019a) (GSCAN consortium), N = 941 280 | European | ALSPAC Cohort 2500 | European | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 37 | Andrews et al. (2020) | Liu et al. (2019a) (GSCAN consortium), N = 537 349; Sanchez-Roige et al. (2019), N = 121 604; Walters et al. (2018) effective-N = 34 780 | European | Lambert et al. (2013) (IGAP) effective-N = 46 668; Huang et al. (2017) effective-N = 37 002 | European | LD r2 < 0.001 | Yes, LD r2 > 0.8 | None |
| 38 | Nishiyama et al. (2019) | Wakai et al. (2011) (The Japan Multi-Institutional Collaborative Cohort (J-MICC) Study), N = 13 618 | Japanese | Wakai et al. (2011) (The Japan Multi-Institutional Collaborative Cohort (J-MICC) Study), N = 13 618 | Japanese | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 39 | Almeida et al. (2014b) | HIMS Cohort, N = 3873 | Predominantly European | HIMS cohort, N = 3873 | Predominantly European | n.a. (1 SNP) | No | None |
| 40 | Wium-Andersen et al. (2015a) | The Copenhagen General Population Study (CGPS) Cohort, N = 78 154 | European | CGPS Cohort, N = 78 154 | European | LD r2 < 0.01 | No | Bonferroni |
| 41 | Polimanti et al. (2019) | Wray et al. (2018) effective-N = 389 039; Walters et al. (2018) effective-N = 30 053 (note that only unrelated individuals were selected); UKB alcohol use frequency – N = 438 308, alcohol use quantity – N = 307 098 | European | Wray et al. (2018) effective-N = 389 039; Walters et al. (2018) effective-N = 30 053 (note that only unrelated individuals were selected); UKB alcohol use frequency – N = 438 308, alcohol use quantity – N = 307 098 | European | LD r2 < 0.01 | No | Bonferroni |
| 42 | Zhou et al. (2020) | Howard et al. (2019) effective-N = 684 817; Schizophrenia working group of PGC, 2014 effective-N = 111 487; Stahl et al. (2019) effective-N = 49 367; Nagel et al. (2018) Neuroticism – N = 449 484, depressed affect, – N = 357 957, worry – N = 348 219; Jansen et al. (2019) – N = 1.3 million; Lee et al. (2018) without UKB effective-N = 179 185 (based on the first paragraph of the results section MVP phase1, effective-N = 114 847 + MVP phase 2, effective-N = 37 485 + PGC effective-N = 26 853) | Predominantly European (small Asian cohorts in schiz GWAS) | Howard et al. (2019) effective-N = 684 817; Schizophrenia working group of PGC, 2014 effective-N = 111 487; Stahl et al. (2019) effective-N = 49 367; Nagel et al. (2018) Neuroticism – N = 449 484, depressed affect – N = 357 957, worry – N = 348 219; Jansen et al. (2019) – N = 1.3 million; Lee et al. (2018) without UKB effective-N = 179 185 (based on the first paragraph of the results section MVP phase1, effective-N = 114 847 + MVP phase2, effective-N = 37 485 + PGC effective-N = 26 853) | Predominantly European (small Asian cohorts in schiz GWAS) | Independent SNPs as reported in exposure GWAS were selected | Yes, LD r2 > 0.8 | Bonferonni |
| 43 | Irons et al. (2007) | McGue et al. (2007) (The Sibling Interaction and Behavior Study), N = 180 | East Asian (Korean) | McGue et al. (2007) (The Sibling Interaction and Behavior Study), N = 180 | East Asian (Korean) | n.a. (1 SNP) | No | None |
| 44 | Chao et al. (2017) | The BeTwiSt project (adolescents from Beijing), N = 1608 | Chinese | The BeTwiSt project (adolescents from Beijing), N = 1608 | Chinese | n.a. (1 SNP) | No | None |
| 45 | Hodgson et al. (2020) | Stringer et al. (2016) (International Cannabis Consortium) effective-N = 31 933; Wray et al. (2018) effective-N = 374 559 | European | Stringer et al. (2016) (International Cannabis Consortium) effective-N = 31 933; Wray et al. (2018) effective-N = 374 559 | European | LD r2 < 0.1 | No | p < 0.01 considered significant |
| 46 | Soler Artigas et al. (2019) | Demontis et al. (2019 effective-N = 49 017; Stringer et al. (2016) (International Cannabis Consortium) effective-N = 31 933 | European | Demontis et al. (2019 effective-N = 49 017; Stringer et al. (2016) (International Cannabis Consortium) effective-N = 31 933 | European | LD r2 < 0.05 | No | None |
| 47 | Pasman et al. (2018) | Schizophrenia working group of PGC, 2014 effective-N = 111 486; International Cannabis Consortium effective-N = 180 934 | Predominantly European (small Asian cohorts in schiz GWAS) | Schizophrenia working group of PGC, 2014 effective-N = 111 486; International Cannabis Consortium effective-N = 180 934 | Predominantly European (small Asian cohorts in schiz GWAS) | LD r2 < 0.001 | Yes, LD r2 ≥ 0.8 | None |
| 48 | Vaucher et al. (2018) | Stringer et al., 2016 (International Cannabis Consortium) effective-N = 31 933 | European | Schizophrenia working group of PGC, 2014 effective-N = 78 227 (note that this sample size is lower than that of the original GWAS, and it was not stated how this subsample was selected) | European (unclear whether the Asian cohorts were included) | 10 leading SNPs (not genome-wide significant) from the exposure GWAS, no criteria for independence stated | No | None |
| 49 | Gage et al. (2017a) | Stringer et al. (2016) (International Cannabis Consortium) effective-N = 31 933; Schizophrenia working group of PGC 2014 effective-N = 111 486 | Predominantly European (small Asian cohorts in schiz GWAS) | Stringer et al. (2016) (International Cannabis Consortium) effective-N = 31 933; Schizophrenia working group of PGC 2014 effective-N = 111 486 | Predominantly European (small Asian cohorts in schiz GWAS) | r2 < 0.9 (LD was corrected for with correlation matrix) | Yes, LD r2 > 0.9 | None |
| 50 | Zhou et al. (2018) | Meta-analysis of 10 European cohorts (the 1958 British birth cohort (1958BC), UKB, Mothers of Avon Longitudinal Study of Parents and Children (ALSPAC-M), Northern Finland Birth Cohorts 1966 (NFBC1966), Cardiovascular Risk in Young Finns Study (YFS), Helsinki Birth Cohort Study (HBCS), Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS), Uppsala Longitudinal Study of Adult Men (ULSAM), Swedish twin registry (STR), and TwinGene), N = 415 530 (of which 300 760 coffee consumers) | European | Meta-analysis of 10 European cohorts (1958BC, ALSPAC-M, NFBC1966, YFS, HBCS, PIVUS, ULSAM, STR, and TwinGene), N = 415 530 (of which 300 760 coffee consumers) | European | Independent SNPs as reported in exposure GWAS were selected | Yes, LD r2 1.0 and 0.97 for two SNPs | None |
| 51 | Treur et al. (2018) | Cornelis et al. (2015) (Caffeine Genetics Consortium) – N = 91 462; Cornelis et al. (2016) – N = 9876; Jones et al. (2016) – N = 128 266; Hammerschlag et al. (2017) effective-N = 92 415 | European | Cornelis et al. (2015) (Caffeine Genetics Consortium), N = 91 462; Cornelis et al. (2016), N = 9876; Jones et al. (2016), N = 128 266; Hammerschlag et al. (2017) effective-N = 92 415 | European | LD r2 < 0.001 | Yes, LD r ≥ 0.8 | None |
| 52 | Kwok et al. (2016) | Cornelis et al. (2015) (Caffeine Genetics Consortium) – N = 129 788 (note that the reported sample size implies that the trans-ethnic data was used) | Predominantly European (~6% African American) | Major Depressive Disorder Working Group of PGC, 2013 effective-N = 18 755; Lambert et al. (2013) effective-N = 46 668 | European | Independent SNPs as reported in exposure GWAS were selected | Yes, LD r2 > 0.8 | Bonferroni |
| 53 | Ding et al. (2019) | Health and Retirement Study (HRS) – N = 3935 | Not clearly indicated (representative of >50 year-olds in the US) | Health and Retirement Study (HRS) N = 3935 | Not clearly indicated (representative of >50 year-olds in the US) | Independent SNPs as reported in exposure GWAS were selected, after which some were excluded to prevent horizontal pleiotropy | No | None |
| 54a | Yuan et al. (2020a) | Lee et al. (2018), N = 1 131 881; Savage et al. (2018), N = 269 867 | European | Liu et al. (2019a) (GSCAN consortium) alcohol drinks per week, N = 941 280; age at onset smoking, N = 341 427; cigarettes per day, N = 337 334 | European | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 55 | Davies et al. (2019) | Individual-level MR: UKB, N = 93 135; Summary-level MR: Okbay et al. (2016), N = 293 723; Hill et al. (2019), N = 248 723 | European | Individual-level MR: UKB, N = 93 135 Summary-level MR: UKB smoking initiation effective-N = 136 760, current smoking effective-N = 46 573 | European | Independent SNPs as reported in exposure GWAS were selected, further clumped with LD r2 < 0.01 | No | None |
| 56a | Davies et al. (2018a) | Individual-level MR: UKB – N = 315 436; Summary-level MR: Okbay et al. (2016) – N = 293 723 | European | Individual-level MR: UKB – N = 315 436; Summary-level MR: not clear from the manuscript, note that the summary-level analyses were used as a sensitivity analysis to check for pleiotropy and were not the main aim | European | LD r2 < 0.001 | Yes, LD r2 = 1 | None |
| 57 | Harrison et al. (2020b) | Individual-level MR: UKB N = 336 997 [N = 336 997/2 (split sample) for lifetime smoking]; Summary-level MR: Liu et al. (2019a) (GSCAN consortium) without UKB and 23andme smoking initiation – N = 249 171, alcohol drinks per week – N = 226 223; UKB Lifetime smoking – N = 336 997 (N = 336 997/2 (split sample)) | European | Individual-level MR: UKB – N = 336 997; Summary-level MR: UKB Lifetime smoking – N = 336 997 [N = 336 997/2 (split sample)] | European | Independent SNPs as reported in exposure GWAS were selected | Yes, LD r2 > 0.8 | Bonferonni |
| 58 | Mahedy et al. (2021) | Liu et al. (2019a) (GSCAN consortium) effective-N = 1 220 901; Pasman et al. (2018) (International Cannabis Consortium) – N = 184 765 | European | ALSPAC Cohort, N = 3232 | European | Independent SNPs as reported in exposure GWAS were selected | No | None |
| 59 | Andrews et al. (2021) | Sanchez-Roige et al. (2019) – N = 141 932; Liu et al. (2019a) (GSCAN consortium) smoking initiation effective-N = 1 220 901, alcohol drinks per week – N = 941 280, cigarettes per day – N = 337 334 | European | Lambert et al. (2013) effective-N = 46 670; Kunkle et al. (2019) effective-N = 57 692; Huang et al. (2017) – N = 40 255 | European | LD r2 < 0.001 | Yes, LD r2 ⩾ 0.8 | FDR |
| 60 | Larsson et al. (2017) | Thorgeirsson et al. (2010) (TAG consortium) smoking initiation effective-N = 72 710, cigarettes per day N = 38 181, smoking cessation effective-N = 41 278;. Jorgenson et al. (2017), N = 71 071; Cornelis et al. (2015), N = 91 462 | European | Lambert et al. (2013) effective-N = 46 668 | European | LD r2 < 0.2 | Yes, LD r2 > 0.9 | Bonferroni (0.05/24 test = 0.002), <0.05 was reported as ‘suggestive evidence’) |
| 61 | Wootton et al. (2020) | Abdellaoui et al. (2019) – N = 511 280; Liu et al. (2019a, 2019b) (GSCAN consortium) without UKB and without 23 and me smoking initiation effective-N = 244 920, alcohol drinks per week – N = 226 223; Walters et al. (2018) effective-N = 34 780 | European | Abdellaoui et al. (2019) – N = 511 280; Liu et al. (2019a, 2019b) (GSCAN consortium) without UKB and without 23andme smoking initiation effective-N = 244 920, cigarettes per day – N = 249 171, smoking cessation effective-N = 142 612, alcohol drinks per week – N = 226 223; Walters et al. (2018) effective-N = 34 780 | European | Independent SNPs as reported in exposure GWAS were selected; for instruments at threshold p < 1 × 10−5, LD r2 < 0.01 | Yes, LD r2 ≥ 0.8 | None |
| 62 | Lim et al. (2020) | Stringer et al. (2016) (International Cannabis Consortium) effective-N = 31 933; Walters et al. (2018), effective-N = 34 780 | European | UKB – N = 125 742 | European | LD r2 < 0.001 | No | None |
| 63 | Treur et al. (2019) | Liu et al. (2019a) (GSCAN consortium) smoking initiation effective-N = 1 220 901, alcohol drinks per week – N = 941 280; Sanchez-Roige et al. (2019) – N = 121 604; Walters et al. (2018) effective-N = 34 779; Pasman et al. (2018) (International Cannabis Consortium) effective-N = 180 934; Demontis et al. (2019) effective-N = 49 017; Cornelis et al. (2016) – N = 91 462 | European | Liu et al. (2019a) (GSCAN consortium) excluding 23andme smoking initiation effective-N = 632 783, cigarettes per day – N = 263 954, smoking cessation – N = 312 821, alcohol drinks per week – N = 537 341; Sanchez-Roige et al. (2019) – N = 121 604; Walters et al. (2018) effective-N = 34 779; Pasman et al. (2018) (International Cannabis Consortium) effective-N = 180 934; Demontis et al. (2019) – N = 15 548 (only adults included); Cornelis et al. (2016) – N = 91 462 | European | Independent SNPs as reported in exposure GWAS were selected | No | None (note the authors explain how they define strength of evidence) |
a
Pre-print publication (not peer-reviewed) obtained from bioRxiv.org, medRxiv.org or arXiv.org.
Note that the complete references to the samples listed under ‘GWAS sample exposure variable(s)’ and ‘GWAS sample outcome variable(s)’ can be found in the original publications (1–63).