Skip to main content
NCBI home page

This is a preprint.

It has not yet been peer reviewed by a journal.

The National Library of Medicine is running a pilot to include preprints that result from research funded by NIH in PMC and PubMed.

Research Square logoLink to Research Square
[Preprint]. 2024 Jul 16:rs.3.rs-3955955. [Version 1] doi: 10.21203/rs.3.rs-3955955/v1

Genome-wide meta-analyses of cross substance use disorders in European, African, and Latino ancestry populations

Dongbing Lai 1, Michael Zhang 2, Nick Green 3, Marco Abreu 4, Tae-Hwi Schwantes-An 5, Clarissa Parker 6, Shanshan Zhang 7, Fulai Jin 8, Anna Sun 9, Pengyue Zhang 10, Howard Edenberg 11, Yunlong Liu 12, Tatiana Foroud
PMCID: PMC11275984  PMID: 39070649

Abstract

Genetic risks for substance use disorders (SUDs) are due to both SUD-specific and SUD-shared genes. We performed the largest multivariate analyses to date to search for SUD-shared genes using samples of European (EA), African (AA), and Latino (LA) ancestries. By focusing on variants having cross-SUD and cross-ancestry concordant effects, we identified 45 loci. Through gene-based analyses, gene mapping, and gene prioritization, we identified 250 SUD-shared genes. These genes are highly expressed in amygdala, cortex, hippocampus, hypothalamus, and thalamus, primarily in neuronal cells. Cross-SUD concordant variants explained ~ 50% of the heritability of each SUD in EA. The top 5% individuals having the highest polygenic scores were approximately twice as likely to have SUDs as others in EA and LA. Polygenic scores had higher predictability in females than in males in EA. Using real-world data, we identified five drugs targeting identified SUD-shared genes that may be repurposed to treat SUDs.

Introduction

Substance use disorders (SUDs, including alcohol, cannabis, opioid, and nicotine) have devasting consequences on individuals, their families, and the society. Globally, approximately 5.5% of disability-adjusted life-years are attributable to SUDs1. For each SUD, twin and family studies have shown that genetic factors are responsible for ~ 50% of the variation2. Searching for SUD-associated genes will not only help us understand the genetic etiologies of SUDs, but it can also facilitate the development of novel prevention and treatment strategies by identifying drugs targeting genes related to SUDs.

SUDs share common features such as uncontrolled use of substances, withdrawal/negative affect, and compromised executive functioning3. Many people use more than one substance and suffer from multiple SUDs simultaneously4. This comorbidity among SUDs suggested shared genetic components across different SUDs, as demonstrated by twin studies5–7 and genetic correlation studies8,9. Furthermore, recent large-scale genome-wide association studies (GWAS) of individual SUD have identified genes associated with more than one SUD, e.g., DRD210,11, FTO10,12–14, and PDE4B10,11,13. To identify SUD-shared genes, one way is to perform univariate analysis by defining SUDs cases as having any SUD and defining controls as not having any SUD. However, individual level genotype data are typically required and thus studies using these methods have small to moderate samples sizes15–19, and as a result, their findings remain to be validated. Multivariate methods such as meta-analysis and genomic structural equation modeling (genomic SEM)20 can utilize summary statistics from large-scale GWAS and thereby are more powerful. Recently, large-scale studies using genomic SEM to search for SUD-shared genes reported novel associations8,21. However, they only explained small portion of genetic variations of SUDs and many SUD-shared genes remain to be discovered8,21.

To identify more SUD-shared genes, we performed the largest cross-SUD meta-analyses to date. GWAS summary statistics of problematic alcohol use (PAU)10,12,22, cannabis use disorder (CUD)23, opioid use disorder (OUD)24,25, and nicotine use disorder (NUD)26,27 were included. We first performed cross-SUD meta-analyses in European (EA), African (AA), and Latino ancestry (LA) samples. Since many individuals have multiple SUDs simultaneously, it is unlikely that a variant increases the risk for one SUD but decreases the risk for another SUD in the same person. Therefore, we only retained variants that had the same directions of effects in different SUDs (i.e., SUD-concordant variants). We also performed three cross-ancestry meta-analyses in EA and AA (EA_AA), EA and LA (EA_LA), EA, AA, and LA (EA_AA_LA) samples. If a variant is associated with SUDs in different populations, then it likely has the same direction of effect across different populations; therefore, we only retained population-concordant variants in cross-ancestry meta-analyses. These SUDs are correlated and thus we used the fixed-effects meta-analysis instead of genomic SEM as it is more powerful28–30. We performed gene-based analyses using SUD-concordant variants, limited to those genes that are expressed in at least one of the 13 brain regions in GTEx31. We used multiple approaches to map variants to genes and prioritize mapped genes. Then for prioritized genes, we performed brain dissection and cell-type enrichment analyses using the single-cell RNA expression data from the BRAIN Initiative Cell Census Network (BICCN)32. Heritability estimates and genetic correlation analyses were also performed. We generated polygenic scores and tested their predictabilities for SUDs using data from the All of Us research program and the Indiana Biobank33. Lastly, we identified drugs targeting prioritized genes then tested whether these drugs could be repurposed to treat SUDs by using real-world data.

Results

Sample description.

GWAS summary statistics included in this study are summarized in Table 1. EA GWAS included PAU10, CUD23, OUD24,25, NUD26,27 and substance abuse from the FinnGen consortium34. Diagnosis of each SUD in these GWAS was based on DSM-IV, ICD 9/10 codes, problematic subscale of the Alcohol Use Disorder Identification Test, and the Fagerstrom Test for Nicotine Dependence. SUDs cases in FinnGen were defined as having any SUD, and controls were those without any SUDs and mental disorders. Sample sizes in EA GWAS ranged from 251,534 to 435,563. AA GWAS included alcohol use disorder (AUD)12,22, CUD23, OUD24,25, and NUD27. Sample sizes ranged from 9,745 to 91,026. LA GWAS included AUD (N = 14,175)12 and OUD (N = 34,861)24. For each SUD, if there was more than one GWAS (e.g., EA OUD GWAS), then they were meta-analyzed first before performing cross-SUD meta-analyses. Sample overlapping across different studies was corrected during meta-analysis. The effective sample sizes after correcting for overlapping samples were 1,467,929 (EA), 159,000 (AA), 45,727 (LA), and 1,672,656 (cross-ancestry total). Using EA samples, the genetic correlations among PAU, OUD, CUD and SUD ranged from 0.48 to 0.70 (P-values ≤ 5.19E-12, Supplemental Table 1), indicating shared genetic underpinning across SUDs.

Table 1.

SUD GWAS included in meta-analyses.

Population Trait Diagnosis # Samples Authors PMID
EA Problematic alcohol use DSM-IV, ICD9/10, AUDIT-P 435,563 Zhou et al 2020 32451486
Opioid use disorder DSM-IV, ICD9/10 308,733 Polimanti et al 2020, Kember et al 2022 32099098, 36171425
Cannabis use disorder DSM-IV, ICD9/10 374,287 Johnson et al 2020 33096046
Nicotine use disorder ICD10, FTND 291,103 Watanabe et al 2019, Quach et al 2020 31427789, 33144568
FinnGen Substance abuse ICD9/10 251,534 NA NA
Total 1,467,929
AA Alcohol use disorder DSM-IV, ICD9/10 62,928 Walters et al 2018, Kranzler et al 2019 30482948, 30940813
Opioid use disorder DSM-IV, ICD9/10 91,026 Polimanti et al 2020, Kember et al 2022 32099098, 36171425
Cannabis use disorder DSM-IV, ICD9/10 9,745 Johnson et al 2020 33096046
Nicotine use disorder FTND 11,787 Quach et al 2020 33144568
Total 159,000
LA Alcohol use disorder ICD9/10 14,175 Kranzler et al 2019 30940813
Opioid use disorder ICD9/10 34,861 Kember et al 2020 36171425
Total 45,757
Cross ancestry total 1,672,656
Open in a new tab

Note: the total samples sizes are effective sample sizes, i.e., samples sizes after correcting sample overlapping. PMID: PubMed ID; AUDIT-P: problematic subscale of the Alcohol Use Disorder Identification Test; FTND: Fagerstrom Test for Nicotine Dependence. AA: African ancestry. EA: European ancestry. LA: Latino ancestry.

Identification of loci associated with SUDs.

We defined independent lead variants as those genome-wide significant (GWS) variants with the smallest P-values and linkage disequilibrium (LD) r2 < 0.1 with other GWS variants. We defined a significant locus as a chromosome region surrounding an independent lead variant bordered by variants having LD r2 > 0.6 with the independent lead variant. If the distance between two loci was < 250 kb, then they were merged. In cross-SUD meta-analyses, the numbers of independent lead variants identified were: 68 (40 loci) in EA, 1 (1 locus) in AA, and 4 (3 loci) in LA (Supplemental Tables 2–4). However, the locus in AA was an intergenic variant without any LD support; therefore, it was likely a false positive and was excluded from further analysis. rs1229984 in ADH1B, which is a well-known alcohol metabolism gene that is associated with AUD35, was identified in LA. LA had only AUD and OUD GWAS and it is likely that this finding was driven by AUD, but we included this locus in subsequent analyses as we cannot rule out its role in other SUDs. In cross-ancestry meta-analysis, the numbers of independent lead variants identified were: 29 (20 loci) in AA_EA, 35 (24 loci) in EA_LA, and 17 (12 loci) in AA_EA_LA (Supplemental Tables 5–7). Manhattan plots of cross-SUD and cross-ancestry meta-analyses are in Supplemental Fig. 1. By merging loci identified in all meta-analyses that have inter-loci distances < 250 kb, we identified 45 loci in total (Table 2). LocusZoom plots of these 45 loci in different meta-analyses are in Supplemental Fig. 2. Among them, 29 loci were identified by multiple meta-analyses (Table 2); and many of them have different independent lead variants in different meta-analyses (Supplemental Fig. 2).

Table 2.

Genome-wide significant loci and their lead variants.

Locus Lead variant Chr BP Effect Allele Non-Effect Allele Z P value Population Locus start Locus end Additional population
1 rs4233254 1 29,168,054 T C 5.627 1.83E-08 EA 28,634,681 29,169,593 EA_LA
2 rs6690101 1 66,414,039 T C -7.891 3.00E-15 EA 66,334,518 66,547,212 AA EA, EA_LA, AA_EA_LA
3 rs2340403 1 73,835,777 T C −6.775 1.24E-11 AA_EA 73,766,431 74,002,104 EA, EA LA, AA_EA_LA
4 rs66994942 2 22,577,783 A G 6.55 5.76E-11 EA 22,430,795 22,971,097 AA_EA
5 rs472140 2 45,139,904 T C 6.591 4.36E-11 EA_LA 45,121,466 45,175,585 EA, AA_EA
6 rs11309539 2 58,025,739 T TA 5.576 2.46E-08 AA_EA 57,942,987 58,444,610 EA
7 rs10865306 2 58,913,139 T C −5.51 3.59E-08 EA 58,863,670 58,932,674
8 rs4625930 2 101,249,256 A G −5.813 6.14E-09 EA 101,232,647 101,317,319 EA_LA
9 rs35942385 2 144,208,523 T G −5.792 6.94E-09 EA_LA 144,145,478 144,263,280 EA
10 rs9812360 3 49,402,351 T C 6.601 4.09E-11 AA_EA 48,724,599 49,890,967 EA, EA LA, AA_EA_LA
11 rs1694933 3 84,951,716 A C −5.996 2.02E-09 AA_EA 84,853,161 84,951,716 EA
12 rs62250713 3 85,513,793 A G 5.75 8.95E-09 EA 85,403,982 85,671,909 EA_LA
13 rs6824152 4 46,915,161 T C −5.618 1.93E-08 AA_EA_LA 46,879,127 46,972,899 AA_EA
14 rs1229984 4 100,239,319 T C −5.512 3.55E-08 LA 100,239,319 100,239,319
15 rs13109404 4 102,896,591 T G 6 1.97E-09 EA_LA 102,702,364 103,001,649 EA
16 rs72678864 4 112,422,145 A G −5.841 5.19E-09 EA 112,303,764 112,503,872 EA_LA
17 rs4413540 5 30,814,680 T C 5.468 4.54E-08 EA 30,813,302 30,842,054
18 rs71627581 5 43,161,351 A G −6.69 2.23E-11 AA_EA 43,125,795 43,190,647 EA
19 rs40506 5 60,525,210 T C 5.592 2.24E-08 EA 60,484,179 60,614,879
20 rs61258990 6 19,278,344 CT C −5.668 1.45E-08 EA 19,211,776 19,358,341
21 rs73377013 6 22,200,119 T C 5.771 7.87E-09 LA 21,318,867 22,702,294
22 rs115133207 6 24,394,925 T C 5.492 3.98E-08 LA 24,355,002 24,427,763
23 rs1796520 6 26,410,800 T C −6.529 6.63E-11 EA_LA 26,364,111 26,577,186 EA
24 rs4713121 6 27,722,064 T C 5.578 2.44E-08 EA_LA 27,347,808 27,980,220 EA, AA_EA
25 rs6467958 7 75,622,281 T C −6.242 4.33E-10 EA 75,607,155 75,852,480 AA EA, EA_LA, AA_EA_LA
26 rs2189010 7 114,119,430 A G 7.433 1.06E-13 EA 113,981,217 114,290,415 AA EA, EA_LA, AA_EA_LA
27 rs10264648 7 114,956,168 T C 5.496 3.88E-08 EA 114,940,159 115,025,708
28 rs2551777 7 135,100,476 T C −6.075 1.24E-09 EA_LA 135,050,259 135,221,170 EA
29 rs72671424 8 93,056,264 A G 6.501 7.97E-11 EA 92,976,563 93,180,965
30 rs12345339 9 12,369,424 T C −5.642 1.69E-08 EA 12,358,209 12,432,665
31 rs405039 9 17,076,950 A C 5.778 7.55E-09 EA_LA 17,044,774 17,127,924 EA
32 rs12380310 9 127,973,473 A G −6.533 6.47E-11 EA 127,780,246 128,008,537 EA_LA
33 rs7044246 9 134,866,271 T C −5.509 3.61E-08 EA 134,802,170 134,946,805
34 rs1591660 10 110,500,212 T C −5.514 3.50E-08 EA_LA 110,462,847 110,635,222
35 rs7476 11 46,342,834 A C −5.641 1.69E-08 EA 46,336,995 46,705,193
36 rs11229119 11 57,535,966 T C −6.179 6.44E-10 EA 57,409,538 57,756,568
37 rs1116313 11 113,296,107 A G 8.473 2.39E-17 AA_EA_LA 112,826,867 113,692,660 EA, AA_EA,EA_LA
38 rs647905 11 121,534,938 T C 6.215 5.14E-10 EA_LA 121,495,141 121,553,314 EA, AA_EA,AA_EA_LA
39 rs11432080 13 96,970,919 G GT −5.645 1.65E-08 EA 96,901,225 97,029,792
40 rs6576007 14 104,323,777 T G −8.329 8.18E-17 EA 103,991,478 104,363,528 AA EA, EA_LA, AA_EA_LA
41 rs12908163 15 47,641,522 T C 7.726 1.11E-14 EA 47,613,403 47,685,416 AA EA, EA_LA, AA_EA_LA
42 rs12232639 18 26,570,703 A G −5.525 3.29E-08 EA 26,570,584 26,611,564 AA_EA,AA_EA_LA
43 rs2613765 19 5,066,330 A G −5.705 1.17E-08 AA_EA 5,046,070 5,086,979 EA
44 rs62128800 19 10,667,750 T C 5.776 7.64E-09 AA_EA 10,648,933 10,686,769
45 rs1231281 19 49,239,200 A G 6.098 1.08E-09 EA_LA 49,206,108 49,259,529 EA
Open in a new tab

Note: Chr: chromosome; BP: base pair positions. AA: African ancestry. EA: European ancestry. LA: Latino ancestry.

There are 104 genes that have at least one GWS variant within its boundaries (transcription start and end sites ± 1 kb, Supplemental Tables 2–7). Seven variants change protein products (missense or stop gain): rs11604671 in ANKK1, rs11601425 in TMPRSS5, rs601338 in FUT2, rs2287922 in RASIP1, rs1229984 in ADH1B, rs3736781 in BTN1A1, and rs1057868 in POR. ANKK1 and TMPRSS5 are ~ 8 kb and ~ 212 kb from DRD2, a well-known gene associated with SUDs. ANKK1 is a part of signal transduction pathway and TMPRSS5 is a member of the serine protease family. It is noteworthy that the missense variant rs11604671 in ANKK1 exhibited P-values < 0.05 in all SUD GWAS in EA. FUT2 and RASIP1 are ~ 50 kb and ~ 25 kb from FGF21, which is related to taste liking measurement and alcohol consumption. FUT2 encodes the alactoside 2-L-fucosyltransferase enzyme and RASIP1 is related to GTPase binding and protein homodimerization activities. BTN1A1 belongs to immunoglobulin superfamily. POR is reported as related to coffee and tea consumption in the GWAS catalog36. Additionally, there are 10 protein changing variants in nine genes that have LD r2 > 0.6 with independent lead variants (Supplemental Table 8; all had P-values ≤ 1.85E-06). Seven of them are in ANKK1, BTN1A1, FUT2, POR, RASIP1, and TMPRSS5 but the other three are in different genes (rs61785974 in PHACTR4, rs6720 in MDH2, rs589292 in SCAI). All these genes are reported in the GWAS catalog as associated with SUD-related or neuropsychiatric-related traits36.

We only required variants having concordant effects by design; however, some GWS variants also showed certain degree of associations (e.g., P-value < 0.05) across all SUDs. In EA, there were 302 GWS variants with P-value < 0.05 in all five EA SUD GWAS (Supplemental Table 2). Among these, 169 are located in 25 genes, with most of them (ANKK1, ARIH2, DRD2, IHO1, KDM4B, LINC01360, NCAM1, PDE4B, PDE4B-AS1, PPP1R13B-DT, PRKAR2A, QRICH1, RABEPK, RUNX1T1, SLC25A20, and USP19) reported as SUD-associated in the GWAS catalog36. No GWS variants had P-value < 0.05 in all cross-ancestry GWAS.

Gene-based analysis.

MAGMA37 was used to perform gene-based analysis within each ancestral population, and gene-based results from each ancestry population were meta-analyzed also using MAGMA37. After Bonferroni correction, the numbers of significant genes identified were: 124 in EA, 1 in AA, 2 in LA, 137 in AA_EA, 109 in EA_LA, and 137 in AA_EA_LA (Supplement tables 9–14). In total, 169 unique genes were identified by gene-based analysis. Manhattan plots of gene-based analyses are in Supplemental Fig. 3.

Mapping significant variants to genes.

For all GWS variants, we used positional mapping (whether a GWS variant is in a gene), eQTL mapping (whether a GWS variant is a cis-eQTL of a gene) and chromatin interaction mapping (whether a GWS variant interacts with a gene) to identify genes impacted by each variant. Similar to gene-based analysis, we limited to genes that are expressed in at least one brain tissue from GTEx31. The numbers of genes identified were: 217 in EA, 0 in AA, 6 in LA, 127 in AA_EA, 148 in EA_LA, and 85 in AA_EA_LA (Supplemental Tables 15–19). In total, 244 unique genes were identified.

Gene prioritization.

The combination of gene-based analysis and gene mapping identified a total of 299 genes. However, some of them may be simply close to SUD-associated genes but not SUD related; therefore, we used two strategies to prioritize mapped genes. First, we checked which genes were in any of nine SUD-related pathways (alcoholism, amphetamine addiction, cocaine addiction, morphine addiction, nicotine addiction, dopaminergic synapse, GABAergic synapse, glutamatergic synapse, and MAPK signaling) from the Kyoto Encyclopedia of Genes and Genomics (KEGG: https://www.genome.jp/kegg/); and nine genes are within these pathways. We also prioritized genes in the same gene families as those in SUDs pathways (12 genes), or that directly interact with genes in SUDs pathways (73 genes) using the STRING database38. In total, this strategy prioritized 94 genes (Supplemental Table 20). Second, we identified 240 genes in the GWAS catalog36 that are associated with any psychiatric trait, brain measurement, and brain function. Together, these two strategies prioritized 250 genes, which include 220 protein coding genes, 15 non-coding RNA, 10 antisense genes, and 5 pseudogenes, (Supplemental Table 20). Some genes are worth noting. For instance, OPRD1, which was identified by positional mapping, is a member of opioid receptor signaling pathway and was nominated as SUD-related in some small studies39–41, but was not reported as SUD-related in the GWAS catalog36.

Brain dissection and cell type enrichment analyses.

For the 250 genes prioritized, we investigated in which brain dissections and cell types that they were significantly highly or lowly expressed. High-throughput single-cell RNA sequencing data from the Human Brain Cell Atlas v1.0 generated by the BRAIN Initiative cell census network32 were used. This data sampled 105 dissections from 10 brain regions across the forebrain, midbrain, and hindbrain, and identified 461 cell types. Prioritized genes were highly expressed in 35 dissections with 1 from amygdala, 27 from cortex, 2 from hippocampus, 3 from hypothalamus, and 2 from thalamus (Supplemental Table 21). Prioritized genes were lowly expressed in 2 dissections (one from pons and another from medulla, Supplemental Table 22). Prioritized genes were highly expressed in 125 of the 461 cell types, the majority of which are neuronal cells (Supplemental Table 23). Most of these cell types are from amygdala, basal forebrain, cerebral cortex, hippocampus, hypothalamus, and thalamus, in agreement with what were observed in the dissection enrichment analysis. Prioritized genes were lowly expressed in one cell type (splatter) from midbrain (Supplemental Table 24).

Heritability estimation and heritability explained by SUD-concordant variants.

Due to the complicated LD patterns and mismatches of LD structures between the AA and LA samples and external LD reference panels, heritability estimation was performed only in EA samples. We identified 477,686 SUD-concordant variants in EA and using them, the estimated SNP heritability (h2snp) of SUDs (using LD score regression42) was 0.10 (SE = 0.001). SUD-concordant variants explained 46.6%, 53.0%, 59.1%, and 50.8% of h2snp of PAU, OUD, CUD and NUD in EA (Supplemental Table 25).

Genetic correlations.

We performed genetic correlation analyses using 1,151 GWAS summary statistics available at CTG-VL (https://vl.genoma.io/data). We identified 344 significant correlations after Bonferroni correction (Supplemental Table 26). Among them, 20 were related substance use, 44 were related to psychiatric diseases, 48 were related behavior, and 9 were related to life events. It is worth noting that while current alcohol drinker status was negatively correlated with SUDs (rG=−0.31, P-value = 3.09E-05), previous alcohol drinker status was positively correlated with SUDs (rG = 0.83, P-value = 1.64E-19).

Polygenic scores (PGS) analyses.

We calculated PGS and tested their predictability in samples from the All of Us research program (Version 7.1) and the Indiana Biobank33. Since one of the major goals of PGS analyses is to identify high-risk individuals, we compared those top 5% individual with the highest PGS to everyone else. Additionally, since males and females have different prevalence of SUDs, we also performed sex-stratified analysis. In EA, the top 5% individuals were about two times more likely to have SUDs compared to everyone else in both All of Us (odds ratio (OR) = 2.21, 95% confidence interval (CI): 2.03–2.39) and the Indiana Biobank (OR = 1.88, 95%CI: 1.50–2.35), and PGS had larger ORs in females (All of US: OR = 2.35, 95%CI: 2.08–2.65; Indiana Biobank: OR = 2.03, 95%CI: 1.50–2.76) than in males (All of Us: OR = 2.07, 95%: 1.85–2.32; Indiana Biobank: OR = 1.71, 95%: 1.23–2.39) (Table 3). In LA, Indiana Biobank did not have sufficient sample size and thus PGS analyses were conducted in All of Us only. Those top 5% individuals were also approximately twice as likely to have SUDs (OR = 1.99, 95%CI: 1.71–2.33), but PGS had similar ORs in females (OR = 1.98, 95%CI: 1.65–2.51) and males (OR = 1.99, 95%CI: 1.62–2.44). In AA, PGS were not significant in any analysis (Table 3).

Table 3.

PGS analyses results.

Sample Population Sex TotalN Total N.case(%) Total N.control(%) Top5% N Top5% N.case(%) Top5% N.control(%) Estimate StdError OR (95%CI) P-value
AOU AA Both 47,516 6,090 (12.82%) 41,426 (87.18%) 2,303 292 (12.68%) 2,011 (87.32%) −0.01 0.07 0.99 (0.87–1.13) 0.92
AOU AA Male 20,989 3,699 (17.62%) 17,290 (82.38%) 1,009 181 (17.94%) 828 (82.06%) 0.00 0.09 1.00 (0.84–1.19) 0.98
AOU AA Female 26,527 2,391 (9.01%) 24,136 (90.99%) 1,294 111 (8.58%) 1,183 (91.42%) −0.08 0.11 0.93 (0.75–1.14) 0.48
AOU EA Both 116,027 7,672 (6.61%) 108,355 (93.39%) 5,710 792 (13.87%) 4,918 (86.13%) 0.79 0.04 2.21 (2.03–2.39) 5.5580
AOU EA Male 45,919 4,405 (9.59%) 41,514 (90.41%) 2,365 451 (19.07%) 1,914 (80.93%) 0.73 0.06 2.07 (1.85–2.32) 5.3937
AOU EA Female 70,108 3,267 (4.66%) 66,841 (95.34%) 3,345 341 (10.19%) 3,004 (89.81%) 0.86 0.06 2.35 (2.08–2.65) 1.0443
AOU LA Both 36,368 2,461 (6.77%) 33,907 (93.23%) 1,818 217 (11.94%) 1,601 (88.06%) 0.69 0.08 1.99 (1.71–2.33) 3.2018
AOU LA Male 12,358 1,543 (12.49%) 10,815 (87.51%) 646 134 (20.74%) 512 (79.26%) 0.69 0.1 1.99 (1.62–2.44) 3.5411
AOU LA Female 24,010 918 (3.82%) 23,092 (96.18%) 1,172 83 (7.08%) 1,089 (92.92%) 0.68 0.12 1.98 (1.56–2.51) 2.2008
IB AA Both 1,829 912 (49.86%) 917 (50.14%) 158 77 (48.73%) 81 (51.27%) −0.01 0.2 0.99 (0.66–1.47) 0.96
IB AA Male 647 371 (57.34%) 276 (42.66%) 54 28 (51.85%) 26 (48.15%) −0.02 0.36 0.98 (0.48–1.98) 0.95
IB AA Female 1,182 541 (45.77%) 641 (54.23%) 104 49 (47.12%) 55 (52.88%) −0.02 0.25 0.98 (0.60–1.61) 0.95
IB EA Both 6,375 2,801 (43.94%) 3,574 (56.06%) 406 248 (61.08%) 158 (38.92%) 0.63 0.11 1.88 (1.50–2.35) 3.3708
IB EA Male 2,862 1,301 (45.46%) 1,561 (54.54%) 180 110 (61.11%) 70 (38.89%) 0.54 0.17 1.71 (1.23–2.39) 1.5603
IB EA Female 3,513 1,500 (42.70%) 2,013 (57.30%) 226 138 (61.06%) 88 (38.94%) 0.71 0.15 2.03 (1.50–2.76) 4.4106
Open in a new tab

Note: N.case: number of cases; N.control: number of controls. Significant P-values are in bold. AOU: All of Us. IB: Indiana Biobank. AA: African ancestry. EA: European ancestry. LA: Latino ancestry.

Drug repurposing.

We queried the Drug Gene Interaction Database (DGIdb, v4.2.0)43 and identified 233 drugs approved by the Food and Drug Administration (FDA, targeting 29 genes; 278 drug-gene pairs). Among them, 93 drugs belong to the nerve system class of the Anatomical Therapeutic Chemical (ATC) Classification System (ATC code N, 159 drug-gene pairs, targeting 10 genes: ANKK1, BDNF, CHRM4, DRD2, GABRA4, HTR3B, NCAM1, OPRD1, PDE4B, and POR) (Supplemental Table 27). To identify potential repurposable drugs, we focused on drugs that: 1) have ATC code N, 2) are not approved to treat any SUD, 3) are shown to be related to SUDs treatment through literature search44–49, and 4) have comparators (i.e., drugs with the same ATC level 4 codes as drugs we identified but not targeting any gene we identified). Five drugs met these criteria: Topiramate in ATC N03AX (treating epilepsy, targeting GABRA4), Desipramine, Imipramine, and Nortriptyline in ATC N06AA (treating depression and anxiety, targeting BDNF) and Methylphenidate in ATC N06BA (treating ADHD, targeting DRD2). We used a large-scale real-world data from the Clinformatics® (https://www.optum.com/business/life-sciences/commercial-analytics/managed-markets.html; sample sizes 325,542 to 1,817,258; Supplemental Table 28) to investigate whether these five drugs reduced the risk for SUDs. Compared to users of comparators, the hazard ratios for developing SUDs were: 0.44 for users of Topiramate (95%CI: 0.42–0.47), 0.89 (95% CI: 0.84–0.94) for users of Desipramine, Imipramine, or Nortriptyline; and 0.84 (95% CI: 0.78–0.91) for users of Methylphenidate, indicating that these five drugs significantly reduce the risk for SUDs and may be repurposed to treat SUDs.

Discussion

In this study, using SUD-concordant and population-concordant variants, we identified 45 loci associated with SUDs in EA, AA, LA, and cross-ancestry meta-analyses. Through gene-based analyses, gene mapping, and gene prioritization, we identified 250 genes associated with SUDs. These genes are highly expressed in amygdala, cortex, hippocampus, hypothalamus, and thalamus, primarily in neuronal cells. SUD-concordant variants explained about 50% of heritability of each SUD in EA. The top 5% EA and LA individuals with the highest PGS calculated from SUD-concordant variants were about twice as likely to have SUDs compared to everyone else in their groups. PGS had larger odds ratios in females than in males in EA. We identified five FDA approved drugs to treat other diseases but could potentially be repurposed to treat SUDs.

Identifying SUD-shared genes is challenging. Each SUD is diagnosed by using multiple criteria and thus is heterogenous, and recent large-scale GWAS had demonstrated the highly polygenic nature of each SUD10,12,22–27. To search for SUD-shared gene, we combined all SUDs together as one phenotype thereby making it more heterogenous and more polygenic, consequently, an even larger sample size is needed. Furthermore, in large-scale SUD GWAS, some SUD-specific variants had extremely small P-values, and their P-values could remain genome-wide significant in multivariate analyses and often they were mistakenly identified as SUD-shared. Additionally, many variants irrelevant to SUDs (i.e., noises) could have small P-values (although not genome-wide significant) simply due to random variations. They usually were included in calculating PGS and thus decreased the PGS predictability. We conducted the largest multivariate analyses to date; and although we focused on concordant variants only, we identified more loci than previous studies8,21. Most importantly, by focusing on SUD-concordant variants, many SUD-specific variants and noises were excluded. This strategy was validated through the out-of-sample PGS analyses using two independent samples as the PGS had high predictability in EA and LA.

In our study, we replicated 14 of the 17 loci reported by Hatoum et al. in their multivariate analysis using the genomic SEM8. Of the 14 replicated loci, 11 had concordant effects and were genome-wide significant, three had concordant effects but were not genome-wide significant (rs2860846, EA P-value = 5.70E-06; rs55855024, EA P-value = 2.21E-06; and rs2861190, EA P-value = 1.88E-04). The three loci failed to replicate in our analyses had opposite directions of effects in some SUDs: rs1260326 (EA P-value = 2.07E-08, NUD had different direction), rs2424952 (EA P-value = 6.20E-04, CUD had different direction), and rs28567725 (EA P-value = 2.55E-07; AA P-value = 0.11; CUD and NUD had different directions of effects in both EA and AA). We also observed similar genetic correlations as those reported in Hatoum et al.,8 but we identified more significant genetic correlations and observed higher magnitude of estimated genetic correlations, likely due to the larger sample sizes used in this study. Overall, we confirmed most previous findings. The different results across studies are likely due to the different strategies used (meta-analyses and retaining concordant variants VS genomic SEM8), different phenotypes included (e.g., nicotine use disorder VS problematic tobacco use8), and larger sample size in our study.

Koob and Volkow proposed a heuristic framework of SUDs based on animal studies and human brain imaging data3. The framework involves three brain regions related to three stages of SUDs: basal ganglia (binge/intoxication stage), extended amygdala (withdrawal/negative affect stage), and prefrontal cortex (preoccupation/anticipation stage)3. We found that prioritized genes were highly expressed in prefrontal cortex, showing an overlap with the proposed framework. Furthermore, we found that among all 34 dissections from cortex, prioritized genes were highly expressed in 27 of them, warranting additional studies to characterize the role of cortex in SUDs. We did not find that the prioritized genes were highly expressed in any basal nuclei dissection after multiple testing correction, although caudate and putamen had unadjusted P-values 0.02 and 0.08, respectively. Prioritized genes were highly expressed in one amygdala dissection but not in extended amygdala. Note, while prioritized genes were not highly expressed in basal nuclei and extended amygdala, they were not lowly expressed in these two regions either. Therefore, they are not unrelated to SUDs but may play less important roles in SUDs than other regions. We found that prioritized genes were highly expressed mostly in neuronal cells, indicating that expressions of SUDs related genes are also cell type specific. Further studies are needed to confirm our findings.

In EA and LA, we found that PGS had odds ratios around 2, which is the suggested threshold for PGS to be incorporated in clinical settings to provide additional information for the risk evaluation50. For those at high-risk, early interventions could be applied to reduce the incident of SUDs, and personalized treatment could be developed to improve the probability of recovery. Additionally, PGS can be measured at any time, even before the initiation of risky substance use or risky behavior. All of these make it tempting to use PGS alone to identify high-risk individuals; however, we want to caution readers of problems of only using PGS. First, the estimated heritability of SUDs was ~ 50% and PGS only explained a portion of heritability; therefore, only using PGS will generate unreliable risk estimation. Second, unlike many diseases, the effect of PGS for SUDs could be significantly reduced or amplified by environmental factors across varying contexts. Some individuals having higher PGS may not use any substance due to their religions, cultures, and health conditions, or simply due to not having access to licit or illicit substances; therefore, PGS is meaningless for these individuals. On the other hand, some individuals with lower PGS may develop SUDs as a result of traumatic life events or psychiatric conditions and using PGS alone will overlook these individuals. To reliably predict the risk for SUDs, both PGS and environmental factors, as well as their interactions should be carefully considered.

We used a large-scale real-world data and a robust active comparator design to identify repurposable drugs. For those five drugs we identified, while previous studies have shown that they could be used to treat one or a few SUDs44–49; we showed that they could potentially be used to treat all SUDs as they are targeting SUD-shared genes. It is worth noting that four of the five drugs we identified (Desipramine, Imipramine, Nortriptyline, and Methylphenidate) target BDNF and DRD2–two genes are also the targets of drugs approved by FDA to treat SUDs such as Bupropion, Disulfiram, and Methadone. This may make them easier to get approval due to the well-understood mechanisms and these drugs should be prioritized for testing their efficacy in clinical trials.

There were limitations to this study. First, the sample sizes of both AA and LA discovery and target datasets are small, resulting in limited statistical power. Additionally in LA, we only have GWAS of AUD and OUD. Second, due to complicated LD patterns and different admixtures in AA and LA, we cannot accurately estimate heritability and genetic correlations in these populations. Third, we used fixed effect meta-analysis, which assumes the effect sizes are the same across different studies. This is a very strong assumption, especially for cross-ancestry analyses. Fourth, some variants may have different directions of effects for different SUDs and/or in different populations and they were filtered out by our strategy. Fifth, real-world data-based analysis were subject to unmeasured confounding and inconsistency between health insurance records and true health statues.

As addressed by de Hemptinne and Posthuma, genetic findings should not be taken out of context, misinterpreted, or misused51. In this study, we identified more variants in EA than in AA and LA, and the predictability of PGS were higher in EA than in AA and LA. We emphasis that these differences are due to much smaller sample sizes in AA and LA as well as the lack of appropriate statistical methods for admixed populations such as AA and LA. These differences do not suggest that AA and LA have different genetic mechanisms of SUDs. Rather, our results point to the critical and urgent need to increase the sample sizes of under-represented populations to reduce health disparities. We also emphasis that genetic factors or PGS (as an overall measure of effects of genetic factors) are not determination for SUD status. Genetic factors and environment factors such as traumatic life events and external stresses contribute about equally to SUDs. Findings from genetic studies must be interpreted with caution and cannot be used to predict SUD-related outcomes without carefully considering environment factors. Our findings should never be used to discriminate or to stigmatize people, or to deny access to prevention and treatment programs, especially for those from vulnerable populations. Our findings should only be used for research purposes to promote health.

In conclusion, using SUD-concordant and population-concordant variants, we have identified multiple genes associated with SUDs. These findings can help us elucidate the mechanisms of SUDs. PGS calculated using SUD-concordant variants had higher predictability in EA and LA. We also identified five drugs targeting genes associated with SUDs that may be repurposed to treat SUDs.

Acknowledgments

The All of Us Research Program is supported by the National Institutes of Health, Office of the Director: Regional Medical Centers: 1 OT2 OD026549; 1 OT2 OD026554; 1 OT2 OD026557; 1 OT2 OD026556; 1 OT2 OD026550; 1 OT2 OD 026552; 1 OT2 OD026553; 1 OT2 OD026548; 1 OT2 OD026551; 1 OT2 OD026555; IAA #: AOD 16037; Federally Qualified Health Centers: HHSN 263201600085U; Data and Research Center: 5 U2C OD023196; Biobank: 1 U24 OD023121; The Participant Center: U24 OD023176; Participant Technology Systems Center: 1 U24 OD023163; Communications and Engagement: 3 OT2 OD023205; 3 OT2 OD023206; and Community Partners: 1 OT2 OD025277; 3 OT2 OD025315; 1 OT2 OD025337; 1 OT2 OD025276. In addition, the All of Us Research Program would not be possible without the partnership of its participants.

We want to acknowledge the participants and investigators of the FinnGen study. The FinnGen study is a large-scale genomics initiative that has analyzed over 500,000 Finnish biobank samples and correlated genetic variation with health data to understand disease mechanisms and predispositions. The project is a collaboration between research organizations and biobanks within Finland and international industry partners.

This study was made possible, in part, with support from the Indiana Clinical and Translational Sciences Institute funded, in part by Award Number UL1TR002529 from the National Institutes of Health, National Center for Advancing Translational Sciences, Clinical and Translational Sciences Award, and the National Center for Research Resources, Construction grant number RR020128 and the Lilly Endowment. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors acknowledge the Indiana University Pervasive Technology Institute for providing [HPC (Big Red II, Karst, Carbonate), visualization, database, storage, or consulting] resources that have contributed to the research results reported within this paper.

Footnotes

Competing interests

The authors declare no competing interests.

Contributor Information

Dongbing Lai, Department of Medical and Molecular Genetics, Indiana University School of Medicine.

Michael Zhang, Indiana University School of Medicine.

Nick Green, Indiana University School of Medicine.

Marco Abreu, Indiana University.

Tae-Hwi Schwantes-An, Department of Medical and Molecular Genetics, Indiana University School of Medicine.

Clarissa Parker, Middlebury College.

Shanshan Zhang, Case Western Reserve University.

Fulai Jin, Case Western Reserve University.

Anna Sun, Indiana University School of Medicine.

Pengyue Zhang, Indiana University.

Howard Edenberg, Indiana University School of Medicine.

Yunlong Liu, Indiana University School of Medicine.

Data availability

EA problematic alcohol use GWAS (Zhou et al 2020), EA, AA, and LA opioid use disorder GWAS (Kember et al 2022), AA alcohol use disorder (Kranzler et al 2019) are available through dbGaP: https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001672. EA and AA opioid use disorder GWAS (Polimanti et al 2020): https://figshare.com/articles/dataset/sud2020-op/14672211. EA and AA cannabis use disorder GWAS (Johnson et al 2020): https://figshare.com/articles/dataset/sud2020-cud/14842692. EA nicotine use disorder GWAS (Watanabe et al 2019): https://atlas.ctglab.nl/ukb2_sumstats/41204_F17_logistic.EUR.sumstats.MACfilt.txt.gz. EA and AA nicotine use disorder GWAS (Quach et al 2022) are available through dbGaP: https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001532.v2.p1. FinnGen Substance abuse: https://www.finngen.fi/en/access_results. AA alcohol use disorder (Walters et al 2018): https://figshare.com/articles/dataset/sud2018-alc/14672187. The Human Brain Cell Atlas v1.0 generated by BRAIN Initiative cell census network: https://cellxgene.cziscience.com/collections/283d65eb-dd53-496d-adb7-7570c7caa443. All of Us research program: https://www.researchallofus.org/data-tools/workbench/. Indiana Biobank: https://indianabiobank.org/. Indiana biobank substance use disorder cohort is also available from dbGaP: https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs003025.v1.p1. All GWAS summary statistics from this study will be deposited to GWAS catalog: https://www.ebi.ac.uk/gwas/. All variants and their weights used to calculated PGS will be available from PGS catalog: https://www.pgscatalog.org/.

References

  • 1.Degenhardt L., et al. The global burden of disease attributable to alcohol and drug use in 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. The Lancet Psychiatry 5, 987–1012 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Deak J.D. & Johnson E.C. Genetics of substance use disorders: a review. Psychol Med 51, 2189–2200 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Koob G.F. & Volkow N.D. Neurobiology of addiction: a neurocircuitry analysis. Lancet Psychiatry 3, 760–773 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Bhalla I.P., Stefanovics E.A. & Rosenheck R.A. Clinical Epidemiology of Single Versus Multiple Substance Use Disorders: Polysubstance Use Disorder. Medical Care 55(2017). [DOI] [PubMed] [Google Scholar]
  • 5.Kendler K.S., Jacobson K.C., Prescott C.A. & Neale M.C. Specificity of Genetic and Environmental Risk Factors for Use and Abuse/Dependence of Cannabis, Cocaine, Hallucinogens, Sedatives, Stimulants, and Opiates in Male Twins. American Journal of Psychiatry 160, 687–695 (2003). [DOI] [PubMed] [Google Scholar]
  • 6.Kendler K.S., Myers J. & Prescott C.A. Specificity of Genetic and Environmental Risk Factors for Symptoms of Cannabis, Cocaine, Alcohol, Caffeine, and Nicotine Dependence. Arch Gen Psychiat 64, 1313–1320 (2007). [DOI] [PubMed] [Google Scholar]
  • 7.Kendler K.S., et al. Recent advances in the genetic epidemiology and molecular genetics of substance use disorders. Nature Neuroscience 15, 181–189 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Hatoum A.S., et al. Multivariate genome-wide association meta-analysis of over 1 million subjects identifies loci underlying multiple substance use disorders. Nature Mental Health 1, 210–223 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Hatoum A.S., et al. The addiction risk factor: A unitary genetic vulnerability characterizes substance use disorders and their associations with common correlates. Neuropsychopharmacology 47, 1739–1745 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Zhou H., et al. Genome-wide meta-analysis of problematic alcohol use in 435,563 individuals yields insights into biology and relationships with other traits. Nature Neuroscience 23, 809–818 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Xu H., et al. Identifying genetic loci and phenomic associations of substance use traits: A multi-trait analysis of GWAS (MTAG) study. Addiction (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Kranzler H.R., et al. Genome-wide association study of alcohol consumption and use disorder in 274,424 individuals from multiple populations. Nature Communications 10, 1499 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Deak J.D., et al. Genome-wide association study in individuals of European and African ancestry and multi-trait analysis of opioid use disorder identifies 19 independent genome-wide significant risk loci. Mol Psychiatr (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Gaddis N., et al. Multi-trait genome-wide association study of opioid addiction: OPRM1 and beyond. Sci Rep 12, 16873 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Johnson E.O., et al. KAT2B polymorphism identified for drug abuse in African Americans with regulatory links to drug abuse pathways in human prefrontal cortex. Addict Biol 21, 1217–1232 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Xiang B., Yang B.Z., Zhou H., Kranzler H. & Gelernter J. GWAS and network analysis of co-occurring nicotine and alcohol dependence identifies significantly associated alleles and network. Am J Med Genet B Neuropsychiatr Genet 180, 3–11 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Wetherill L., et al. Genome-wide association study identifies loci associated with liability to alcohol and drug dependence that is associated with variability in reward-related ventral striatum activity in African- and European-Americans. Genes Brain Behav 18, e12580 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Chang S.H., et al. Genome-wide association meta-analyses identify novel genetic risk loci and polygenic phenotype associations for heroin, methamphetamine and alcohol dependences. Clin Transl Med 12, e659 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Zhang H., et al. Strong and weak cross-inheritance of substance use disorders in a nationally representative sample. Mol Psychiatr 27, 1742–1753 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Grotzinger A.D., et al. Genomic structural equation modelling provides insights into the multivariate genetic architecture of complex traits. Nature Human Behaviour 3, 513–525 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Schoeler T., et al. Novel Biological Insights Into the Common Heritable Liability to Substance Involvement: A Multivariate Genome-wide Association Study. Biol Psychiat 93, 524–535 (2023). [Google Scholar]
  • 22.Walters R.K., et al. Transancestral GWAS of alcohol dependence reveals common genetic underpinnings with psychiatric disorders. Nat Neurosci 21, 1656–1669 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Johnson E.C., et al. A large-scale genome-wide association study meta-analysis of cannabis use disorder. Lancet Psychiatry 7, 1032–1045 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Kember R.L., et al. Cross-ancestry meta-analysis of opioid use disorder uncovers novel loci with predominant effects in brain regions associated with addiction. Nature Neuroscience 25, 1279–1287 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Polimanti R., et al. Leveraging genome-wide data to investigate differences between opioid use vs. opioid dependence in 41,176 individuals from the Psychiatric Genomics Consortium. Mol Psychiatry 25, 1673–1687 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Watanabe K., et al. A global overview of pleiotropy and genetic architecture in complex traits. Nat Genet 51, 1339–1348 (2019). [DOI] [PubMed] [Google Scholar]
  • 27.Quach B.C., et al. Expanding the genetic architecture of nicotine dependence and its shared genetics with multiple traits. Nat Commun 11, 5562 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Solovieff N., Cotsapas C., Lee P.H., Purcell S.M. & Smoller J.W. Pleiotropy in complex traits: challenges and strategies. Nature Reviews Genetics 14, 483–495 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Zhu Z., Anttila V., Smoller J.W. & Lee P.H. Statistical power and utility of meta-analysis methods for cross-phenotype genome-wide association studies. PLOS ONE 13, e0193256 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Lee P.H., et al. Genomic Relationships, Novel Loci, and Pleiotropic Mechanisms across Eight Psychiatric Disorders. Cell 179, 1469–1482.e1411 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Lonsdale J., et al. The Genotype-Tissue Expression (GTEx) project. Nature Genetics 45, 580–585 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Siletti K., et al. Transcriptomic diversity of cell types across the adult human brain. Science 382, eadd7046 (2023). [DOI] [PubMed] [Google Scholar]
  • 33.Lai D., et al. Gene-based polygenic risk scores analysis of alcohol use disorder in African Americans. Transl Psychiatry 12, 266 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Kurki M.I., et al. FinnGen provides genetic insights from a well-phenotyped isolated population. Nature 613, 508–518 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Edenberg H.J. & McClintick J.N. Alcohol Dehydrogenases, Aldehyde Dehydrogenases, and Alcohol Use Disorders: A Critical Review. Alcoholism-Clinical and Experimental Research 42, 2281–2297 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Sollis E., et al. The NHGRI-EBI GWAS Catalog: knowledgebase and deposition resource. Nucleic Acids Res 51, D977–d985 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.de Leeuw C.A., Mooij J.M., Heskes T. & Posthuma D. MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput Biol 11, e1004219 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Szklarczyk D., et al. The STRING database in 2021: customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res 49, D605–d612 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Burns J.A., et al. Molecular Imaging of Opioid and Dopamine Systems: Insights Into the Pharmacogenetics of Opioid Use Disorders. Front Psychiatry 10, 626 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Nelson E.C., et al. Association of OPRD1 polymorphisms with heroin dependence in a large case-control series. Addict Biol 19, 111–121 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Levran O., Randesi M., Adelson M. & Kreek M.J. OPRD1 SNPs associated with opioid addiction are cis-eQTLs for the phosphatase and actin regulator 4 gene, PHACTR4, a mediator of cytoskeletal dynamics. Transl Psychiatry 11, 316 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Finucane H.K., et al. Partitioning heritability by functional annotation using genome-wide association summary statistics. Nature Genetics 47, 1228-+ (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Freshour S.L., et al. Integration of the Drug-Gene Interaction Database (DGIdb 4.0) with open crowdsource efforts. Nucleic Acids Res 49, D1144–D1151 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Kranzler H.R., et al. Topiramate treatment for heavy drinkers: moderation by a GRIK1 polymorphism. Am J Psychiatry 171, 445–452 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Johnson B.A., et al. Topiramate for the treatment of cocaine addiction: a randomized clinical trial. JAMA Psychiatry 70, 1338–1346 (2013). [DOI] [PubMed] [Google Scholar]
  • 46.Kosten T., et al. Desipramine and contingency management for cocaine and opiate dependence in buprenorphine maintained patients. Drug Alcohol Depend 70, 315–325 (2003). [DOI] [PubMed] [Google Scholar]
  • 47.Agabio R., Trogu E. & Pani P.P. Antidepressants for the treatment of people with co-occurring depression and alcohol dependence. Cochrane Database Syst Rev 4, Cd008581 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Prochazka A.V., et al. A randomized trial of nortriptyline for smoking cessation. Arch Intern Med 158, 2035–2039 (1998). [DOI] [PubMed] [Google Scholar]
  • 49.Aryan N., et al. The therapeutic effects of methylphenidate and matrix-methylphenidate on addiction severity, craving, relapse and mental health in the methamphetamine use disorder. Subst Abuse Treat Prev Policy 15, 72 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Hao L., et al. Development of a clinical polygenic risk score assay and reporting workflow. Nat Med (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.de Hemptinne M.C. & Posthuma D. Addressing the ethical and societal challenges posed by genome-wide association studies of behavioral and brain-related traits. Nature Neuroscience 26, 932–941 (2023). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

EA problematic alcohol use GWAS (Zhou et al 2020), EA, AA, and LA opioid use disorder GWAS (Kember et al 2022), AA alcohol use disorder (Kranzler et al 2019) are available through dbGaP: https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001672. EA and AA opioid use disorder GWAS (Polimanti et al 2020): https://figshare.com/articles/dataset/sud2020-op/14672211. EA and AA cannabis use disorder GWAS (Johnson et al 2020): https://figshare.com/articles/dataset/sud2020-cud/14842692. EA nicotine use disorder GWAS (Watanabe et al 2019): https://atlas.ctglab.nl/ukb2_sumstats/41204_F17_logistic.EUR.sumstats.MACfilt.txt.gz. EA and AA nicotine use disorder GWAS (Quach et al 2022) are available through dbGaP: https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001532.v2.p1. FinnGen Substance abuse: https://www.finngen.fi/en/access_results. AA alcohol use disorder (Walters et al 2018): https://figshare.com/articles/dataset/sud2018-alc/14672187. The Human Brain Cell Atlas v1.0 generated by BRAIN Initiative cell census network: https://cellxgene.cziscience.com/collections/283d65eb-dd53-496d-adb7-7570c7caa443. All of Us research program: https://www.researchallofus.org/data-tools/workbench/. Indiana Biobank: https://indianabiobank.org/. Indiana biobank substance use disorder cohort is also available from dbGaP: https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs003025.v1.p1. All GWAS summary statistics from this study will be deposited to GWAS catalog: https://www.ebi.ac.uk/gwas/. All variants and their weights used to calculated PGS will be available from PGS catalog: https://www.pgscatalog.org/.

Articles from Research Square are provided here courtesy of American Journal Experts

RESOURCES