Neuroscience and Biobehavioral Reviews 140 (2022) 104806

Contents lists available at ScienceDirect

Neuroscience and Biobehavioral Reviews

journal homepage: www.elsevier.com/locate/neubiorev

Index of multiple deprivation contributed to common psychiatric disorders:

A systematic review and comprehensive analysis
Xin Qi a, 1, Yumeng Jia b, 1, Chuyu Pan b, 1, Chune Li b, Yan Wen b, Jingcan Hao c, Li Liu b,
Bolun Cheng b, Shiqiang Cheng b, Yao Yao b, Feng Zhang b, *
a
Precision Medicine Center, The First Affiliated Hospital of Xi’an Jiaotong University, Xi’an, China
b
Key Laboratory of Trace Elements and Endemic Diseases of National Health and Family Planning Commission, School of Public Health, Health Science Center, Xi’an
Jiaotong University, Xi’an, China
c
Cancer Center, The First Affiliated Hospital of Xi’an Jiaotong University, Xi’an, China

A R T I C L E I N F O A B S T R A C T

Keywords: Background: Limited studies have been conducted to explore the interaction effects of social environmental and
Index of multiple deprivation genetic factors on the risks of common psychiatric disorders.
Genome-wide gene-environment interaction Methods: 56,613–106,695 individuals were collected from the UK Biobank cohort. Logistic or liner regression
study
models were first used to evaluate the associations of index of multiple deprivation (IMD) with bipolar disorder
Psychiatric disorders
Interactions
(BD), depression and anxiety in UK Biobank cohort. Then, for the significant IMD associated with BD, depression
Social environmental deprivations and anxiety, genome-wide gene-environment interaction study (GWEIS) was performed by PLINK 2.0.
Result: Totally, the higher levels of IMD were significantly associated with higher risks of BD, depression and
anxiety. For BD, GWEIS identified multiple significant SNPs interacting with IMD, such as rs75182167 for in­
come and rs111841503 for education. For depression and anxiety, GWEIS found significant SNPs interacting
with income and education, such as rs147013419 for income and rs142366753 for education.
Conclusion: Social environmental deprivations contributed to the risks of psychiatric disorders. Besides, we re­
ported multiple candidate genetic loci interacting with IMD, providing novel insights into the biological
mechanism.

1. Introduction anxiety were approximately 1 %, 27.2 % and 25 %, respectively (Dis­

eases and Injuries, 2020; Merikangas et al., 2011; Rotenstein et al.,
Compared with communicable diseases, non-communicable diseases 2016). Besides, the individuals with psychiatric disorders had other
had a rising burden to global disease burden, according to Global Burden comorbidities, great economics burden and a low quality of life
of Disease Study 2019 (GBD 2019) (Diseases and Injuries, 2020). (Krishnan, 2005; Greenberg et al., 2015). Psychiatric diseases were the
Mental, neurological, and substance use disorders were the leading major cause of suicides worldwide (Bachmann, 2018) and the rate of BD
cause (Rehm and Shield, 2019; GBD 2017 Disease and Injury Incidence suicide had approximately 20–30 times than that of the general popu­
and Prevalence Collaborators, 2018), and these disorders caused 7% of lation (Miller and Black, 2020).
global disease burden estimated by disability-adjusted life-years Accumulating evidence showed that socioeconomic deprivation had
(DALYs) and 19 % of disability (Rehm and Shield, 2019). Bipolar dis­ great influence on the risks of BD, depression and anxiety (GBD 2017
order (BD), depression and anxiety were three common psychiatric Disease and Injury Incidence and Prevalence Collaborators, 2018;
disorders and the overall pooled crude prevalence of BD, depression and Marangoni et al., 2016; Hossain et al., 2019). The weighted prevalence

Abbreviations: GBD, Global Burden of Disease; DALYs, Disability-adjusted life-years; BD, Bipolar disorder; IMD, Index of multiple deprivation; GWAS, Genome-
wide association study; MDD, Major depression disorder; GWEIS, Genome-wide gene-environment interactions study; PHQ, Patient health questionnaire; GAD,
Generalized anxiety disorder; QC, Quality control; HRC, Haplotype Reference Consortium; PCs, Principal components; HWE, Hardy-Weinberg equilibrium; GO, Gene
ontology; CC, Cellular components; MF, Molecular function; BP, Biological process.
* Correspondence to: School of Public Health, Health Science Center, Xi’an Jiaotong University, No. 76 Yan Ta West Road, Xi’an 710061, China.
E-mail address: [email protected] (F. Zhang).
1
The three authors contributed equally to this work.

https://doi.org/10.1016/j.neubiorev.2022.104806
Received 17 November 2021; Received in revised form 8 April 2022; Accepted 31 July 2022
Available online 1 August 2022
0149-7634/© 2022 Elsevier Ltd. All rights reserved.
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

of mental disorders in the poor socioeconomic status was higher than in the gene-environment interactions. Finally, the functional analysis of
the rich socioeconomic status (Hashmi and Alam, 2021). A historical genes mapped by the significant SNPs was conducted through gene set
cohort study design showed that the individuals who derived from the enrichment analysis.
most deprived areas were more likely to have poorer health outcomes
and be treated with crisis resolution and home treatment (Kingsford and 2. Methods
Webber, 2010). The increased risk of common mental disorders after
traumatic events had a correlation with the poor income and low edu­ 2.1. UK Biobank cohort
cation, and a higher level of education and income could attenuate the
relationships between stress and common mental disorders (Fahey et al., 2.1.1. Ethics approval
2016). Income, socioeconomic deprivation, living environment and Our study sample were collected from the UK Biobank (application
employment deprivation were risky environments for depression 46478). The UK Biobank study was approved by the National Health
(Dowdall et al., 2017; Fernández-Niño et al., 2014). These risky envi­ Service National Research Ethics Service (11/NW/0382). All of the
ronments can contribute to the index of multiple deprivation (IMD), participants have supplied the informed consent to participate the UK
which consists of seven domains of deprivation according to their Biobank study.
respective weights and is the official and overall relative measure of
deprivation (Noble et al., 2019). A logistic regression model was per­ 2.1.2. Study population
formed in 7353 individuals and found that IMD had a significant pred­ The UK Biobank was a large-scale cohort study of ~500,000 in­
ication for depression (Wickham et al., 2014). Although the evidence of dividuals from the United Kingdom, which were all aged 40–69 years
the relationships between socioeconomic and psychiatric diseases were between 2006 and 2010. From over 9.2 million invitations, 503,325
increasing, the detailed biological mechanism of socioeconomic status participants were recruited. Participants used touchscreen tests and
affecting psychiatric disorders was unclear and needed to be further questionnaires as well as nurse-led interviews to self-report the exten­
researched. sive phenotypic data upon baseline assessment. The individuals who had
BD, depression and anxiety were three complex psychiatric diseases IMD were selected in this study.
with strong genetic factors, with the estimated heritability of 70–90 %,
31–42 % and 30 %, respectively (Hashmi and Alam, 2021; Sullivan 2.1.3. Phenotypes definition
et al., 2000; Gordovez and McMahon, 2020). Multiple loci were Three psychiatric disorders were analyzed in this study, including
discovered to be associated with BD, depression and anxiety through BD, depression (including self-reported depression and patient health
genome-wide association study (GWAS) analysis (Stahl and Breen, questionnaire (PHQ)− 9 score) and anxiety (including self-reported
2019; Howard and Adams, 2019; Meier et al., 2019). However, for one anxiety and generalized anxiety disorder (GAD)− 7 score). BD, self-
nucleic acid, the common genetic variants were no less than 5 % chro­ reported depression and self-reported anxiety were defined via self-
mosomes in a population, the effects of which on complex diseases reported according to three UK Biobank fields: 20002 (Non-cancer
usually were very small or none (Levinson et al., 2014). For BD, the only illness code, self-reported), 20126 (Bipolar and major depression status)
approximately 25 % of overall heritability was accounted for those and 20544 (Mental health problems ever diagnosed by a professional).
common variants of risk altogether (Gordovez and McMahon, 2020). PHQ-9 score represented the total score for measuring the depression
Besides, mental diseases were polygenic traits and many genetic variants severity (Kroenke et al., 2010). In UK Biobank, PHQ-9 is based on nine
could influence mental diseases with each of small effect (Ripke et al., depressive symptoms and signs, including Feeling bad about yourself
2013; Zhao et al., 2019a). For the statistical power of discovering the (20507), Trouble concentrating (20508), Feeling down, depressed, or
significant associations, it depended on the sample number of cohorts hopeless (20510), Poor appetite or overeating (20511), Thoughts that
and effect of each SNP. Therefore, the identified genetic variants had you would be better off dead (20513), Little interest or pleasure in doing
limited explanation to the etiology of BD, depression and anxiety. things (20514), Trouble sleeping (20517), Moving or speaking slowly or
Although BD, depression and anxiety were three heritable psychi­ fidgety or restless (20518) and Feeling tired (20519). The all of 9
atric disorders, it was widely recognized that a multifactorial model in symptoms scores (1− 4) in UK Biobank were reduced 1 and then added
which gene and environment interact could best explain the pathogen­ up to meet the total score (0− 27) for PHQ-9. GAD-7 (Kroenke et al.,
esis of these disorders (Craddock and Sklar, 2013). There was accumu­ 2010) was a continuous variable and was used to measure the severity of
lating evidence that the function of most predisposing genes were anxiety with a total score (0− 21). GAD-7 was based on seven anxious
response modifiers to environmental factors but not the primary etiol­ symptoms and signs, including Feeling nervous, anxious or on edge
ogy role in the occurrence of diseases (Tiret, 2002). Through assessing (20506), Not being able to stop or control worrying (20509), Worrying
the occurrence of stressful life events and the onset of major depression too much about different things (20520), Trouble relaxing (20515),
over a 1-year period in female twins, Kenneth et al. found that the odds Being so restless that it is hard to sit still (20516), Becoming easily
ratio was 5.64, 4.52 and 3.58 in all subjects, heterozygous and homo­ annoyed or irritable (20505) and Feeling afraid as if something awful
zygous pairs, respectively (Kendler et al., 1999). Interestingly, an might happen (20512). BD, self-reported depression and self-reported
encouraging result came from a GWEIS in 126,522 genotyped European anxiety are binary variables, while PHQ-9 and GAD-7 score are
individuals, which discovered that the major depression disorder (MDD) continuous variables (Davis and Cullen, 2019). The detailed definitions
with reported trauma exposure had a higher SNP-based heritability than and sample information were provided in the Supplementary materials
MDD without reported trauma exposure (Coleman et al., 2020). Besides, and Table 1.
there was suggestive evidence that gene-environment interactions had
influence in phenotypic abnormalities of BD through structural mag­ 2.1.4. Genotyping, quality control and imputation
netic resonance imaging studies (Geoffroy et al., 2013). Therefore, the For UK Biobank cohort, the detailed genotyping methodology,
research about genome-wide gene-environment interactions study quality control and imputation were descripted in the published studies
(GWEIS) contributes a key role in the understanding of the pathogenesis (Bycroft et al., 2018). In brief, the genotypes of 488,377 participants
of BD, depression and anxiety. were included in the UK Biobank genetic data. Two similar genotyping
In this study, we explored the potential interactions of IMD and SNPs arrays, the Applied Biosystems UK BiLEVE Axiom Array and UK Biobank
for BD, depression and anxiety based on the UK Biobank. First, we Axiom Array, were used to assay genotypes and they shared 95 % of
evaluated the correlations for IMD with BD, depression and anxiety marker content. Genotyping was performed in 106 sequential batches of
using regression models in UK Biobank cohort. Then, for the significant approximately 4700 samples. Quality control (QC) was performed using
and major IMD deprivation domains, GWEIS was performed to access PLINK v1.9 and R v3.3.1. All the subjects included in this study were the

2
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

Table 1
The characteristics of study subjects from UK Biobank.
Bipolar disorder Depression Anxiety

Self-reported depression PHQ-9 score Self-reported anxiety GAD-7 score

Total 56,613 106,695 104,137 94,022 104,643

Cases 2012 52,017 / 19,034 /
Controls 54,601 54,678 / 74,988 /
Sex
Female 27,161 (47.98 %) 60,057 (56.28 %) 58,139 (55.83 %) 51,693 (54.98 %) 58,408 (55.82 %)
Male 29,452 (52.02 %) 46,638 (43.71 %) 45,998 (44.17 %) 42,329 (45.02 %) 46,235 (44.18 %)
Age 57.214 ± 7.499 56.565 ± 7.643 56.356 ± 7.601 56.607 ± 7.555 56.348 ± 7.601
IMD 13.915 ± 11.178 15.790 ± 12.755 14.632 ± 11.679 14.669 ± 11.743 14.646 ± 11.693
Income score 0.090 ± 0.076 0.103 ± 0.087 0.095 ± 0.079 0.095 ± 0.079 0.095 ± 0.079
Employment score 0.073 ± 0.049 0.081 ± 0.056 0.076 ± 0.051 0.076 ± 0.051 0.076 ± 0.051
Health score -0.317 ± 0.822 -0.191 ± 0.846 -0.266 ± 0.828 -0.260 ± 0.828 -0.266 ± 0.828
Education score 11.419 ± 12.893 13.588 ± 14.902 12.175 ± 13.489 12.259 ± 13.594 12.195 ± 13.509
Housing score 20.017 ± 10.078 19.705 ± 10.001 19.857 ± 10.057 19.778 ± 10.035 19.847 ± 10.054
Crime score -0.175 ± 0.757 -0.117 ± 0.766 -0.145 ± 0.762 -0.147 ± 0.764 -0.145 ± 0.762
Living environment score 16.512 ± 13.741 17.503 ± 14.349 17.035 ± 14.067 16.953 ± 14.002 17.043 ± 14.067

Notes: Index of Multiple Deprivation, IMD; Patient Health Questionnaire, PHQ; Generalized Anxiety Disorder, GAD.

self-reported white people. The samples were removed for missing 2.4. Genome-wide gene-environment interactions study (GWEIS)
rate> 0.05, high heterozygosity and sex chromosome karyotypes puta­
tively different from XY or XX. SHAPEIT3 was used to phase on the The significant IMD were further selected for GWEIS. Based on the
autosomes, with the reference panel of the 1000 Genomes Phase 3 UK Biobank cohort, we performed GWEIS for BD, self-reported depres­
dataset. The Haplotype Reference Consortium (HRC) reference panel sion, PHQ-9, self-reported anxiety and GAD-7 by PLINK 2.0 using the
and a merged UK10K/1000 Genomes Phase 3 panel were both used to significant IMD as environmental factors (Chang et al., 2015). Besides,
impute the UK Biobank dataset (Walter and Min, 2015; McCarthy and the high quality SNPs with call rates≥ 0.90, Hardy-Weinberg equilib­
Das, 2016). rium (HWE)≥ 0.001 and minor allele frequencies (MAFs)≥ 0.01 were
included in our analysis. Models were adjusted for individual level de­
mographic characteristics (age and sex) and the top ten population
2.2. Environmental factors structure PCs. The significant level for genome-wide gene-environment
interaction effects was identified at P ≤ 4.55 × 10− 9 (5.0 × 10− 8/11) via
Index of multiple deprivation (IMD, England) was given when par­ Bonferroni correction. Furthermore, the significant SNPs were mapped
ticipants joined the UK Biobank cohort and the field code was 26410. to genes using an online annotation tool, GWAS4d (http://mulinlab.
IMD was a measure of relative deprivation for small areas (Lower Layer tmu.edu.cn/gwas4d) (Huang et al., 2018). “CMplot” R script was
Super Output Areas, LSOAs) used to rank neighborhoods across the UK applied to map Circular Manhattan plots using R software (Version
according the output area in which their postcode was located (Noble 4.0.2).
et al., 2019). IMD was generated based on 37 seven separate indicators,
and was composed of seven domains of deprivation according to their 2.5. Functional gene sets enrichment analysis
respective weights, including Income score (26411, 22.5 %), Employ­
ment score (26412, 22.5 %), Health score (26413, 13.5 %), Education The gene sets enrichment analysis was conducted to explore the
score (26414, 13.5 %), Housing score (26415, 9.3 %), Crime score functional relevance of the identified genes interacting with IMD. The
(26416, 9.3 %) and Living environment score (26417, 9.3 %). Each IMD gene sets enrichment analysis was implemented by GENE2FUNC of
domain is ranked from the least to most deprived. The detailed infor­ FUMA GWAS tool (Functional Mapping and Annotation of Genome-
mation of the IMD and the IMD domains were listed in Table 1. Wide Association Studies, https://fuma.ctglab.nl/) (Watanabe and
Taskesen, 2017). FUMA is an integrative web-based platform using in­
formation from multiple biological resources to facilitate functional
2.3. Statistical analysis
annotation of GWAS results. KEGG and Gene ontology (GO), including
cellular components (CC), biological process (BP) and molecular func­
We used logistic regression model to access the associations of IMD
tion (MF), were represented in the functional gene sets for the identified
with BD, self-reported depression and self-reported anxiety, while linear
genes. Bonferroni correction was used in this analysis to control multiple
regression model for PHQ-9 and GAD-7 score. BD, self-reported
tests and the significant level was adjusted P ≤ 0.05. R (version 4.0.2)
depression, PHQ-9 score, self-reported anxiety and GAD-7 score were
was used to visualize the results.
defined as dependent variables. Covariates included sex, age and top 10
principal components (PCs) of population structure, which was used in
3. Results
all models. First, to evaluate the associations, deciles of IMD (10 groups)
were entered into the regression model as a categorical variable and
3.1. The descriptive characteristics of study samples
decile 1 as the least deprivation was selected as the reference category.
Then, to further explore the associations of specific domains of IMD with
The individuals with self-reported “white British” were included in
psychiatric disorders, each of the 7 domains of IMD (income, employ­
our study. The unrelated subjects were generated with KING software
ment, health, education, housing, crime and living environment) were
(Manichaikul et al., 2010). 56,613–106,695 individuals with IMD, BD,
included into regression models as independent variables after
depression and anxiety data were collected from the UK Biobank cohort.
normalization using scale, with five psychiatric diseases as dependent
The detailed characteristics and sample information were shown in
variables, respectively. All of these statistical analyses were conducted
Table 1.
by R (version 4.0.2). P ≤ 0.01 was selected as the significant level based
on Bonferroni correction approach (0.05/5 = 0.01). GraphPad Prism
8.0.2 was used to visualize the results.

3
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

3.2. Observational analysis results of IMD with BD, depression and score, with the highest beta of 1.460 (SE = 0.056) for PHQ-9 score and
anxiety 0.845 (SE = 0.050) for GAD-7 score. In the logistic model of self-
reported depression and anxiety, the individuals in the most deprived
We discovered significant associations of IMD (deciles) with BD, IMD decile were estimated to have 3.87 (95 %CI, 3.64–4.12) and 2.14
depression and anxiety. The results of the logistic and linear regression (95 %CI, 1.98–2.31) times to be depression and anxiety than in the
models were shown in Fig. 1A and Supplementary table 1. The higher decile 1 of IMD, respectively.
levels of IMD were associated with the higher risk of BD, depression and Fig. 1B and Supplementary table 2 showed the association analysis
anxiety. For BD, the highest adjusted odds ratio (OR) was discovered in results for the seven domains of IMD with BD, depression and anxiety.
IMD decile 10 with 7.25 (95 %CI, 5.91–8.89). The highest beta was For BD, income, health, education and living environment domains
estimated in the most deprivation (IMD decile 10) for PHQ-9 and GAD-7 showed the larger effects (adjusted OR>1.1, P < 0.01). Health and

Fig. 1. Relationships between Index of Multiple Deprivation (IMD) and seven IMD domains for BD, self-reported depression, Patient Health Questionnaire (PHQ)− 9
score, self-reported anxiety and Generalized Anxiety Disorder (GAD)− 7 score. (A) The results of regression models for IMD (deciles) with BD, self-reported
depression, PHQ-9 score, self-reported anxiety and GAD-7 score. (B) The results of regression models for seven domains of IMD with BD, self-reported depres­
sion, PHQ-9 score, self-reported anxiety and GAD-7 score. The IMD domains with red symbols were selected to perform the GWEIS with the corresponding
phenotype. Notes: Genome-wide environmental interaction study, GWEIS; * represents P < 0.05; ** represents P < 0.01; *** represents P < 0.001.

4
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

education domains had the lead adjusted OR for self-reported depression NPNT (P = 7.94 × 10− 13 for education) were detected (Fig. 3A, 3C and
and anxiety. Besides, income and education were the most influential Supplementary table 10–11). However, no significant interactions were
domains for PHQ-9, and education contributed more to GAD-7 score. observed for self-reported depression (Supplementary Figure 1A-B and
Supplementary table 8–9). The top 5 significantly lead SNPs were shown
3.3. Interactions between individual SNPs and the IMD domains in Table 2.
For anxiety GAD− 7 score, we identified 50 SNPs interacting with
For BD, GWEIS identified 2738, 245, 3199 and 919 SNP loci (all education, mapping to 4 genes, such as CDH4 (rs35610560, P = 1.19 ×
P < 4.55 × 10− 9) interacting with income, health, education and living (10− 10) and INPP5F (rs117368191, P = 1.67 × 10− 9) (Figs. 3B, 3C and
environment, respectively (Fig. 2 and Supplementary table 3–7). Then, Supplementary table 14). No significant interactions were found for self-
we used GWAS4d to annotate these identified SNPs which were mapped reported anxiety (Supplementary Figure 1C-D and Supplementary table
to 744, 96, 898 and 320 genes for income, health, education and living 12–13). The top 5 significantly lead SNPs were shown in Table 2.
environment, respectively. For example, rs75182167 (DAB1,
P = 1.33 × 10− 19) for income, rs141817494 (ROBO2, 3.4. Functional gene sets enrichment analysis
P = 3.26 × 10− 14) for health, rs111841503 (CACNA1C,
P = 2.54 × 10− 18) for education and rs12257534 (GRK5, To explore the biological functions of genes interacting with income
P = 5.36 × 10− 17) for living environment were found. The top 5 for BD, we conducted GSEA and found that the identified genes of BD
significantly lead SNPs were shown in Table 2. enriched in multiple brain related GO terms, such as neurogenesis (BP,
For depression PHQ− 9 score, GWEIS identified 147 and 180 SNPs adjusted P = 1.12 × 10− 13), synapse (CC, adjusted P = 1.11 × 10− 21)
(all P < 4.55 × 10− 9) interacting with income and education, mapping and cytoskeletal protein binding (MF, adjusted P = 2.34 × 10− 11).
to 58 and 67 genes, respectively. For example, rs147013419 locating at Moreover, we detected 12 candidate KEGG pathways, including type II
GTDC1 (P = 4.05 × 10− 14 for income), rs116954276 locating at diabetes mellitus (adjusted P = 6.96 × 10− 6) and Fc gamma R mediated
SLC14A2 (P = 8.59 × 10− 13 for income), rs142366753 locating at phagocytosis (adjusted P = 1.97 × 10− 2) (Fig. 4 and Supplementary
B3GLCT (P = 2.80 × 10− 13 for education) and rs115892524 locating at table 15). No significant function results for other gene lists interacting

Fig. 2. The interactions of chromosomal regions and IMD

domains for BD. (A) Chromosomal regions interacting with
four IMD domains for BD. From the center, the first circos
depicts the –log10 P-values of each variant due to double
exposure i.e., the effect of both SNP allele and income. The
second circos shows chromosome density. For center, form
inside to outside, the effects for SNP allele with income,
health, education and living environment are shown. Red
plots represent the P value < 4.55 × 10− 9. (B) The number
of significant SNPs (P < 4.55 × 10− 9) and mapping genes
interacting with four IMD domains for Bipolar disorder.

5
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

Table 2
The top 5 significant interaction effects between each IMD and SNP for Bipolar Disorder, PHQ-9 score and GAD-7 score (P < 4.55 × 10− 9).
Phenotypes IMD domains SNP Chr P Gene
− 21
Bipolar disorder Income rs61912060 12 2.35 × 10 PDE3A
21
rs113188157 12 3.25 × 10− AC125616.1
21
rs28608761 12 3.33 × 10− AC125616.1
19
rs75182167 1 1.33 × 10− DAB1
19
rs34873421 11 2.71 × 10− MRVI1
16
Health rs147352632 5 1.64 × 10− LINC01411
15
rs147724156 7 5.21 × 10− HECW1,HECW1-IT1
14
rs141817494 3 3.26 × 10− ROBO2
14
rs141327492 4 6.87 × 10− INPP4B
14
rs146489535 22 7.66 × 10− LINC01658
18
Education rs111841503 12 2.54 × 10− CACNA1C
18
rs147724156 7 2.74 × 10− HECW1,HECW1-IT1
18
rs116836075 1 8.77 × 10− RYR2
18
rs146733762 4 9.65 × 10− GAK
16
rs148201588 11 1.99 × 10− AC103796.1
17
Living environment rs12257534 10 5.36 × 10− GRK5
16
rs185155151 10 5.33 × 10− CTNNA3
16
rs11198890 10 5.77 × 10− GRK5
16
rs77327911 10 6.20 × 10− CTNNA3
15
rs147156506 4 1.14 × 10− ANAPC4
14
PHQ-9 Income rs147013419 2 4.05 × 10− GTDC1
13
rs183240340 5 4.97 × 10− NIM1K
13
rs116954276 18 8.59 × 10− SLC14A2
12
rs76495481 1 1.88 × 10− AL591885.1
12
rs79709397 6 1.91 × 10− EYA4
13
Education rs142366753 13 2.80 × 10− B3GLCT
13
rs115892524 4 7.94 × 10− NPNT
12
rs79091143 5 1.19 × 10− ABLIM3,AC012613.1
12
rs8069981 17 2.87 × 10− LINC00469
12
rs77453425 17 3.31 × 10− LINC00469
10
GAD-7 Education rs35610560 20 1.19 × 10− CDH4
10
rs35363330 20 1.43 × 10− CDH4
10
rs138937208 2 4.70 × 10− DAPL1
9
rs117368191 10 1.67 × 10− INPP5F,PHACTR2P1
9
rs8069981 17 5.29 × 10− LINC00469

Notes: Chromosome, Chr; Patient Health Questionnaire, PHQ; Generalized Anxiety Disorder, GAD.

with IMD for BD, depression and anxiety. and it was urgently needed to research the biological mechanisms of the
interactions for deprivation and mental diseases.
4. Discussion Four IMD domains were discovered to have positively correlations
with BD through logistic regression model, including income, health,
Index of multiple deprivation, as an environmental factor, has a great education and living environment. A nationwide prospective cohort
influence on human health and has been widely used in health related study found that those with depression and bipolar disorders did not
research (Maier, 2017; Greene et al., 2020). The IMD could be selected have a higher education and poor income outcomes compared with the
as a major predictor of the burden of global disease and injury (Newton general population across the entire work-life span (Hakulinen et al.,
et al., 2015). A large-scale study of GWEIS was performed for IMD with 2019). The children, living in inadequate home environments with
BD, depression and anxiety based on the samples from the UK Biobank parents diagnosed with BD and an insufficient level of support, had an
study. In this study, we identified that IMD had positive associations increased risk of BD (Gantriis et al., 2019).
with BD, depression and anxiety as well as a large number of SNPs and Further GWEIS identified multiple candidate genes interacting with
mapping genes interacting with IMD for BD, depression and anxiety IMD for BD, such as DAB1, GRK and CACNA1C. DAB1 mainly enriched in
through GWEIS analysis. neuro related GO, such as neurogenesis, neuro development and neuro
Deprivation had a direct and significant impact on the quality of life differentiation, which indicated that DAB1 might have an interaction
through exacerbating psychiatric disorders, and inhibiting life satisfac­ with income for BD. The protein encoded by DAB1 gene could regulate
tion (Xia and Ma, 2020). The level of social deprivation at birth was neuronal positioning during the development of the brain through
found to have an association with a higher prevalence of psychiatric interacting with protein kinase pathways (Meyer et al., 2003). In addi­
disorders and the increase of social deprivation (O’Donoghue et al., tion, GRK encodes G protein-coupled receptor kinase 5, which is widely
2016). In our study, using logistic and linear regression models, our expressed in the various brain cortex (Erdtmann-Vourliotis et al., 2001).
results also found that IMD positively correlated with psychiatric dis­ The exposure to poor living environments could cause a significant delay
orders, and the higher level of IMD (deciles), the higher risk for psy­ of the central nervous system and brain cerebral cortex (Narducci et al.,
chiatric disorders. The previous studies suggested that the adolescents 2018). Moreover, the knockdown of GRK5 could cause the impairments
from low socioeconomic status were up to three times more likely to of social behavior in medial prefrontal cortex, which indicted that GRK5
have mental health problems compare with high socioeconomic status contributed to maintaining normal social behavior (Niu et al., 2018).
(Amone-P′ Olak et al., 2009). Moreover, based on socioeconomic char­ CASNA1C had a correlation with the mean thickness of cortical brain
acteristics of area, the mental disorders and distress had dynamic areas through magnetic resonance imaging and the previous study
changes and occurred more commonly in socioeconomic disadvantage showed that education had a subtle but robust distributed effect on
(Enticott et al., 2016). Therefore, consistent with our results, the in­ regional cortical thickness (Smedler et al., 2019; Habeck et al., 2020).
terventions on improving the socioeconomics and broadening health BD had a link with abnormalities of specific cortical thickness and
care could reduce the burden of mental diseases (Barnett et al., 2012) CACNA1C was a candidate gene for bipolar disorder supported by strong

6
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

Fig. 3. The interactions of chromosomal regions and IMD

domains for Patient Health Questionnaire (PHQ)− 9 and
Generalized Anxiety Disorder (GAD)− 7 score. (A) Chro­
mosomal regions interacting with two IMD domains for
PHQ-9 score. From the center, the first circos depicts the
–log10 P-values of each variant due to double exposure i.e.,
the effect of both SNP allele and income. The second circos
shows chromosome density. Red plots represent the P value
< 4.55 × 10− 9. Form inside to outside of the center, the
effects for SNP allele with income and education are
shown. (B) Chromosomal regions interacting with educa­
tion for GAD-7 score. C) The number of significant SNPs
and mapping genes for PHQ-9 and GAD-7 score
(P < 4.55 ×(10− 9).

evidence of a large scale analysis (Sklar et al., 2011; Hanford et al., status compared with high income status (Lorant et al., 2003; Korhonen
2016). Combined with our findings, the most significant genes of DAB1, et al., 2017). There existed a dose-response relationship between income
GRK5 and CACNA1C, respectively interacting with income, living and education, and the risk of depression increased in not being in ed­
environment and education, were suggested to contribute to the devel­ ucation compared with being in education (Lorant et al., 2003; Korho­
opment of BD. nen et al., 2017).
For depression, income, health and education were the most influ­ GWEIS suggested that GTDC1 and SLC14A2 could influence
ential deprivation domains in regression models. Research showed that depression through interacting with income. The disruption of GTDC1
the deprivation of health was associated with depression, which indi­ was associated with neurodevelopment through the patient-specific
cated that enhancement of the primary health care was critical to reduce induced pluripotent stem cells (iPSCs) and neurodevelopment played a
the burden of common psychiatric disorders (Meier et al., 2019). The critical role in depression (Lima-Ojeda et al., 2018; Aksoy et al., 2017).
previous studies showed a higher odds of depression in low income The knockdown of UT-B, widely expressing in brain regions and encoded

7
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

Fig. 4. The significant terms (adjusted

P < 0.05) for genes interacting with income for
BD. The abscissa is rich factor. Rich factor
means the ratio of the number of differential
genes in this term in the number of genome
background genes in this term. The larger the
value of rich factor is, the greater the enrich­
ment degree. The ordinate is the pathway term.
The size of the dot indicates the number of
differential genes under the pathway. The
different color means different adjusted P value.
(A) The top 20 GO terms for genes interacting
with income for BD. (B) The KEGG results for
genes interacting with income for BD.

by SLC14A2 and SLC14A1, could induce urea accumulation, the difficulties in their daily life and are more prone to be anxiety. Besides,
decrease of regional cerebral blood flow and morphological changes, the highly educated individuals may have more abilities and resources
and then resulted into the impairment of synaptic plasticity and to seek the health care information and medical assistance (Almeida-­
depression-like behavior (Levinson et al., 2014; Zhao et al., 2019b). A Filho, 2011).
large national cohort discovered that the low income level could in­ A significant interaction of education and CDH4 was found for anx­
crease the risk of incident chronic kidney disease, and there existed iety. CDH4 (cadherin 4) was discovered to be a candidate gene associ­
significant relationships between depression and chronic kidney disease ated with the brain development through whole-exome sequencing and
(Zhao et al., 2019b; Martin et al., 2014). Therefore, we suggested that chromosomal microarray studies for malformations of cortical devel­
multiple genes interacting with deprivation domains could regulate the opment (Reiss, 2013). Besides, cadherins were indicated to be crucial for
development of depression. the neuros migration into the developing neocortical wall (Wiszniewski
Health and education deprivation were the most influential IMD et al., 2018).
domains for anxiety. The undergraduate students showed a higher level UK Biobank was not representative of the general population, which
of death anxiety than graduate students (Nienaber and Goedereis, existed a healthy volunteer bias. On average, participants recruited in
2015). Combined with our results, the individuals with health depri­ UK Biobank are healthier compared with the general population and live
vation and low education level may lack confidence when facing the in less socioeconomically deprived areas compared to nonparticipants

8
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

(Fry et al., 2017). Therefore, combined with our findings that the higher Third, there were no significant results in accessing the interactions of
levels of deprivation were associated with the higher risk of common SNPs and IMD domains for self-reported depression and self-reported
psychiatric disorders, UK Biobank may underestimate the relationships anxiety. A measurement error through self-reported online question­
between deprivation and common psychiatric disorders. Nonetheless, naires should be paid more attention.
the valid associations of deprivation and common psychiatric diseases In conclusion, our results discovered social environmental depriva­
may be widely generalizable and does not require participants to be tion had significant correlations with psychiatric diseases, and suggested
representative of the population at large (Allen et al., 2012). Besides, that there had the significant interactions of genetic variants and IMD
94.6 % of participants were of white ethnicity and the age of individuals domains on psychiatric diseases through regression models and GWEIS
in UK Biobank were older than the general population, which resulted analysis. Our findings indicated that IMD had a major contribution to
that there were lack of research for individuals in other ages and races the risk of psychiatric disorders and supplied the insights to reducing the
(Fry et al., 2017). Although the UK Biobank is a biases sample, its large burden of psychiatric disorders.
size and heterogeneity of exposure measurements provide valid scien­
tific inferences about associations between exposure and health that can Declaration of Competing Interest
be generalized to other populations (Fry et al., 2017; Allen et al., 2012;
Collins, 2012). Moreover, UK Biobank was a large-scale pop­ The authors have stated that they have no conflict of interest.
ulation-based cohort study and collected extensive data directly from
participants, all of which could ensure a better quality of data and lead Data availability
to cleaner association signals (Fry et al., 2017). Meanwhile, UK Biobank
has been widely used in different fields of medicine and they successfully The UKB data are available through the UK Biobank Access Man­
detected meaningful findings helpful to understand the mechanism of agement System https://www.ukbiobank.ac.uk/. We will return the
diseases and promote public health (Lu et al., 2022; Huang et al., 2022; derived data fields following UKB policy; in due course, they will be
Wendt et al., 2022). available through the UK Biobank Access Management System.
IMD, a LSOA level measure of relative deprivation for small areas,
was used for the individual deprivation in this study, which may lead to Acknowledgments
ecological fallacy. According to previous studies, ecological fallacy is
unavoidable in geographical analysis of the association of socio- We thank the UK Biobank participants and the UK Biobank team for
economic deprivation and disease incidence, and choosing the small­ their work in collecting, processing and disseminating these data for
est geographical unit available allows the measure of deprivation to be analysis. The UK Biobank study was approved by the National Health
closest to the individual level (Lokar et al., 2019). Although, in the Service National Research Ethics Service (11/NW/0382). Our data
previous study, through a genome-wide interaction study in UK Bio­ dictionary ID for application is 46478.
bank, Kenneth et al. used IMD as one of the exposures to research the
interactions between genetic variants and non-genetic risk factors
Funding
influencing COVID-19 severity (Westerman et al., 2022), IMD measure
may miss important indicators such as consumption level of the indi­
This study is supported by the National Natural Scientific Foundation
vidual. However, based on UK Biobank data, we were unable to obtain
of China (81922059) and the Natural Science Basic Research Plan in
indicators of socioeconomic deprivation for further smaller regional
Shaanxi Province of China (2021JCW-08).
units. Therefore, the better measure of individual deprivation data or
validation was needed for further research.
We aimed to explore the potential interacting effects of IMD and Appendix A. Supporting information
susceptibility genes on the risks of BD, depression and anxiety using the
UK Biobank samples. Genome-wide gene-environment interaction study Supplementary data associated with this article can be found in the
will naturally confront some challenges, such as large size of samples, online version at doi:10.1016/j.neubiorev.2022.104806.
the impact of confounding factors and multiple testing problem. In this
study, the major and significant IMD domains were used as the envi­ References
ronmental factors to explore gene-environment interacting effects. Be­
Aksoy, I., Utami, K.H., Winata, C.L., 2017. Personalized genome sequencing coupled
sides, some confounding factors possibly affect the interactions of IMD with iPSC technology identifies GTDC1 as a gene involved in neurodevelopmental
and susceptibility genes for common psychiatric diseases, such as disorders. Hum. Mol. Genet. 26 (2), 367–382.
different age, sex and genetic background of study samples. To minimize Allen, N., Sudlow, C., Downey, P., 2012. UK biobank: current status and what it means
for epidemiology. Health Policy Technol. 1 (3), 123–126.
the effect of these factors, individual level demographic characteristics Almeida-Filho, N., 2011. Higher education and health care in Brazil. Lancet 377 (9781),
(age and sex) and the top ten population structure PCs were used as 1898–1900.
covariates in our analysis. Further, GWEIS was used to research the Almli, L.M., Duncan, R., Feng, H., et al., 2014. Correcting systematic inflation in genetic
association tests that consider interaction effects: application to a genome-wide
gene-environment interactions using the regression model in PLINK 2.0 association study of posttraumatic stress disorder. JAMA Psychiatry 71 (12),
(Chang et al., 2015). GWEIS was an advantageous alternative approach 1392–1399.
capable of detecting gene × environment at risk loci across the genome Amone-P′ Olak, K., Burger, H., Ormel, J., Huisman, M., Verhulst, F.C., Oldehinkel, A.J.,
2009. Socioeconomic position and mental health problems in pre- and early-
without relying on a priori hypotheses and while assessing interactions
adolescents: the TRAILS study. Soc. Psychiatry Psychiatr. Epidemiol. 44 (3),
with multiple environmental factors (Werme et al., 2021; Almli et al., 231–238.
2014). Then, to control the potential impact of multiple testing problem, Bachmann, S., 2018. Epidemiology of suicide and the psychiatric perspective. Int J.
Environ. Res Public Health 15, 7.
based on a strict genome-wide significance threshold, Bonferroni
Barnett, K., Mercer, S.W., Norbury, M., Watt, G., Wyke, S., Guthrie, B., 2012.
correction (P < 5.0 × 10− 8/11) was used to select the significant results Epidemiology of multimorbidity and implications for health care, research, and
to ensure the accuracy of our study results (Hayes, 2013; Curtin and medical education: a cross-sectional study. Lancet 380 (9836), 37–43.
Schulz, 1998). Bycroft, C., Freeman, C., Petkova, D., et al., 2018. The UK Biobank resource with deep
phenotyping and genomic data. Nature 562 (7726), 203–209.
However, there still have some limitations in our study. First, albeit Chang, C.C., Chow, C.C., Tellier, L.C., Vattikuti, S., Purcell, S.M., Lee, J.J., 2015. Second-
with a large sample size, it is hard to completely identify the influence of generation PLINK: rising to the challenge of larger and richer datasets. Gigascience
gene-environment interactions on the psychiatric disorders in this 4, 7.
Coleman, J.R.I., Peyrot, W.J., Purves, K.L., et al., 2020. Genome-wide gene-environment
observational study. Second, due to linkage disequilibrium and hetero­ analyses of major depressive disorder and reported lifetime traumatic experiences in
geneity caused by geographic location, our results may be limited. UK Biobank. Mol. Psychiatry 25 (7), 1430–1446.

9
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

Collins, R., 2012. What makes UK Biobank special? Lancet 379 (9822), 1173–1174. Krishnan, K.R., 2005. Psychiatric and medical comorbidities of bipolar disorder.
Craddock, N., Sklar, P., 2013. Genetics of bipolar disorder. Lancet 381 (9878), Psychosom. Med. 67 (1), 1–8.
1654–1662. Kroenke, K., Spitzer, R.L., Williams, J.B., Löwe, B., 2010. The patient health
Curtin, F., Schulz, P., 1998. Multiple correlations and Bonferroni’s correction. Biol. questionnaire somatic, anxiety, and depressive symptom scales: a systematic review.
Psychiatry 44 (8), 775–777. Gen. Hosp. Psychiatry 32 (4), 345–359.
Davis, K.A.S., Cullen, B., 2019. Indicators of mental disorders in UK Biobank – a Levinson, D.F., Mostafavi, S., Milaneschi, Y., 2014. Genetic studies of major depressive
comparison of approaches. Int. J. Methods Psychiatry Res. 28 (3), e1796. disorder: why are there no genome-wide association study findings and what can we
Diseases, G.B.D., Injuries, C., 2020. Global burden of 369 diseases and injuries in 204 do about it? Biol. Psychiatry 76 (7), 510–512.
countries and territories, 1990-2019: a systematic analysis for the global burden of Lima-Ojeda, J.M., Rupprecht, R., Baghai, T.C., 2018. Neurobiology of depression: a
disease study 2019. Lancet 396 (10258), 1204–1222. neurodevelopmental approach. World J. Biol. Psychiatry 19 (5), 349–359.
Dowdall, N., Ward, C.L., Lund, C., 2017. The association between neighbourhood-level Lokar, K., Zagar, T., Zadnik, V., 2019. Estimation of the ecological fallacy in the
deprivation and depression: evidence from the south african national income geographical analysis of the association of socio-economic deprivation and cancer
dynamics study. BMC Psychiatry 17 (1), 395. incidence. Int. J. Environ. Res. Public Health 16, 3.
Enticott, J.C., Meadows, G.N., Shawyer, F., Inder, B., Patten, S., 2016. Mental disorders Lorant, V., Deliège, D., Eaton, W., Robert, A., Philippot, P., Ansseau, M., 2003.
and distress: associations with demographics, remoteness and socioeconomic Socioeconomic inequalities in depression: a meta-analysis. Am. J. Epidemiol. 157
deprivation of area of residence across Australia. Aust. N. Z. J. Psychiatry 50 (12), (2), 98–112.
1169–1179. Lu et al. 2022, the complete details is Lu Y, Tang H, Huang P, Wang J, Deng P, Li Y,
Erdtmann-Vourliotis, M., Mayer, P., Ammon, S., Riechert, U., Höllt, V., 2001. Zheng J, Weng L. Assessment of causal effects of visceral adipose tissue on risk of
Distribution of G-protein-coupled receptor kinase (GRK) isoforms 2, 3, 5 and 6 cancers: a Mendelian randomization study. Int J Epidemiol. 2022 Feb 26:dyac025.
mRNA in the rat brain. Brain Res. Mol. Brain Res. 95 (1–2), 129–137. doi: 10.1093/ije/dyac025. Epub ahead of print. PMID: 35218345.
Fahey, N., Soni, A., Allison, J., 2016. Education mitigates the relationship of stress and Maier, W., 2017. Indizes Multipler Deprivation zur Analyse regionaler
mental disorders among rural Indian women. Ann. Glob. Health 82 (5), 779–787. Gesundheitsunterschiede in Deutschland. Bundesgesundheitsblatt - Gesundh. -
Fernández-Niño, J.A., Manrique-Espinoza, B.S., Bojorquez-Chapela, I., Salinas- Gesundh. 60 (12), 1403–1412.
Rodríguez, A., 2014. Income inequality, socioeconomic deprivation and depressive Manichaikul, A., Mychaleckyj, J.C., Rich, S.S., Daly, K., Sale, M., Chen, W.M., 2010.
symptoms among older adults in Mexico. PLOS One 9 (9), e108127. Robust relationship inference in genome-wide association studies. Bioinformatics 26
Fry, A., Littlejohns, T.J., Sudlow, C., et al., 2017. Comparison of sociodemographic and (22), 2867–2873.
health-related characteristics of UK biobank participants with those of the general Marangoni, C., Hernandez, M., Faedda, G.L., 2016. The role of environmental exposures
population. Am. J. Epidemiol. 186 (9), 1026–1034. as risk factors for bipolar disorder: a systematic review of longitudinal studies.
Gantriis, D.L., Thorup, A.A.E., Harder, S., et al., 2019. Home visits in the Danish high risk J. Affect Disord. 193, 165–174.
and resilience study - VIA 7: assessment of the home environment of 508 7-year-old Martin, J.L., McLean, G., Park, J., et al., 2014. Impact of socioeconomic deprivation on
children born to parents diagnosed with schizophrenia or bipolar disorder. Acta rate and cause of death in severe mental illness. BMC Psychiatry 14, 261.
Psychiatrica Scand. 140 (2), 126–134. McCarthy, S., Das, S., 2016. A reference panel of 64,976 haplotypes for genotype
GBD 2017 Disease and Injury Incidence and Prevalence Collaborators, 2018. Global, imputation. Nat. Genet. 48, 1279 (The Haplotype Reference Consortium).
regional, and national incidence, prevalence, and years lived with disability for 354 Meier, S.M., Trontti, K., Purves, K.L., 2019. Genetic variants associated with anxiety and
diseases and injuries for 195 countries and territories, 1990-2017: a systematic stress-related disorders: a genome-wide association study and mouse-model study.
analysis for the Global Burden of Disease Study 2017. Lancet 392 (10159), JAMA Psychiatry 76 (9), 924–932.
1789–1858. Merikangas, K.R., Jin, R., He, J.P., 2011. Prevalence and correlates of bipolar spectrum
Geoffroy, P.A., Etain, B., Houenou, J., 2013. Gene x environment interactions in disorder in the world mental health survey initiative. Arch. Gen. Psychiatry 68 (3),
schizophrenia and bipolar disorder: evidence from neuroimaging. Front. Psychiatry 241–251.
4 (136-136). Meyer, G., De Rouvroit, C.L., Goffinet, A.M., Wahle, P., 2003. Disabled-1 mRNA and
Gordovez, F.J.A., McMahon, F.J., 2020. The genetics of bipolar disorder. Mol. Psychiatry protein expression in developing human cortex. Eur. J. Neurosci. 17 (3), 517–525.
25 (3), 544–559. Miller, J.N., Black, D.W., 2020. Bipolar disorder and suicide: a review. Curr. Psychiatry
Greenberg, P.E., Fournier, A.A., Sisitsky, T., Pike, C.T., Kessler, R.C., 2015. The economic Rep. 22 (2), 6.
burden of adults with major depressive disorder in the United States (2005 and Narducci, R., Baroncelli, L., Sansevero, G., 2018. Early impoverished environment delays
2010). J. Clin. Psychiatry 76 (2), 155–162. the maturation of cerebral cortex. Sci. Rep. 8 (1), 1187.
Greene, G., Gartner, A., Farewell, D., et al., 2020. Mental health selection: common Newton, J.N., Briggs, A.D., Murray, C.J., 2015. Changes in health in England, with
mental disorder and migration between multiple states of deprivation in a UK analysis by English regions and areas of deprivation, 1990-2013: a systematic
cohort. BMJ Open 10 (2), e033238. analysis for the global burden of disease study 2013. Lancet 386 (10010),
Habeck, C., Gazes, Y., Razlighi, Q., Stern, Y., 2020. Cortical thickness and its associations 2257–2274.
with age, total cognition and education across the adult lifespan. PLOS One 15 (3), Nienaber, K., Goedereis, E., 2015. Death anxiety and education: a comparison among
e0230298. undergraduate and graduate students. Death Stud. 39 (8), 483–490.
Hakulinen, C., Musliner, K.L., Agerbo, E., 2019. Bipolar disorder and depression in early Niu, B., Liu, P., Shen, M., et al., 2018. GRK5 regulates social behavior via suppression of
adulthood and long-term employment, income, and educational attainment: A mTORC1 signaling in medial prefrontal cortex. Cereb. Cortex 28 (2), 421–432.
nationwide cohort study of 2,390,127 individuals. Depress Anxiety 36 (11), O’Donoghue, B., Roche, E., Lane, A., 2016. Neighbourhood level social deprivation and
1080–1088. the risk of psychotic disorders: a systematic review. Soc. Psychiatry Psychiatr.
Hanford, L.C., Nazarov, A., Hall, G.B., Sassi, R.B., 2016. Cortical thickness in bipolar Epidemiol. 51 (7), 941–950.
disorder: a systematic review. Bipolar Disord. 18 (1), 4–18. Rehm, J., Shield, K.D., 2019. Global burden of disease and the impact of mental and
Hashmi, R., Alam, K., 2021. Prevalence of mental disorders by socioeconomic status in addictive disorders. Curr. Psychiatry Rep. 21 (2), 10.
Australia: a cross-sectional epidemiological study. Am. J. Health Promot. 35 (4), Reiss, F., 2013. Socioeconomic inequalities and mental health problems in children and
533–542. adolescents: a systematic review. Soc. Sci. Med 90, 24–31.
Hayes, B., 2013. Overview of statistical methods for genome-wide association studies Ripke, S., Wray, N.R., Lewis, C.M., 2013. A mega-analysis of genome-wide association
(GWAS). In: Gondro, C., van der Werf, J., Hayes, B. (Eds.), Genome-Wide Association studies for major depressive disorder. Mol. Psychiatry 18 (4), 497–511.
Studies and Genomic Prediction. Humana Press, Totowa, NJ, pp. 149–169. Rotenstein, L.S., Ramos, M.A., Torre, M., 2016. Prevalence of depression, depressive
Hossain, S., Anjum, A., Uddin, M.E., Rahman, M.A., Hossain, M.F., 2019. Impacts of symptoms, and suicidal ideation among medical students: a systematic review and
socio-cultural environment and lifestyle factors on the psychological health of meta-analysis. JAMA 316 (21), 2214–2236.
university students in Bangladesh: a longitudinal study. J. Affect Disord. 256, Sklar, P., Ripke, S., Scott, L.J., et al., 2011. Large-scale genome-wide association analysis
393–403. of bipolar disorder identifies a new susceptibility locus near ODZ4. Nat. Genet 43
Howard, D.M., Adams, M.J., 2019. Genome-wide meta-analysis of depression identifies (10), 977–983.
102 independent variants and highlights the importance of the prefrontal brain Smedler, E., Abé, C., Pålsson, E., Ingvar, M., Landén, M., 2019. CACNA1C polymorphism
regions. Nat. Gen. 22 (3), 343–352. and brain cortical structure in bipolar disorder. J. Psychiatry Neurosci. 45 (1),
Huang, D., Yi, X., Zhang, S., 2018. GWAS4D: multidimensional analysis of context- 182–187.
specific regulatory variant for human complex diseases and traits. Nucleic Acids Res. Stahl, E.A., Breen, G., 2019. Genome-wide association study identifies 30 loci associated
46 (W1), W114–W120. with bipolar disorder. Nat. Gen. 51 (5), 793–803.
Huang, Y., Hui, Q., Gwinn, M., 2022. Interaction between genetics and smoking in Stefan Noble, David McLennan, Michael Noble, et al. The English Indices of Deprivation
determining risk of coronary artery diseases. Genet. Epidemiol. 2019 2019.
Kendler, K.S., Karkowski, L.M., Prescott, C.A., 1999. Causal relationship between Sullivan, P.F., Neale, M.C., Kendler, K.S., 2000. Genetic epidemiology of major
stressful life events and the onset of major depression. Am. J. Psychiatry 156 (6), depression: review and meta-analysis. Am. J. Psychiatry 157 (10), 1552–1562.
837–841. Tiret, L., 2002. Gene-environment interaction: a central concept in multifactorial
Kingsford, R., Webber, M., 2010. Social deprivation and the outcomes of crisis resolution diseases. Proc. Nutr. Soc. 61 (4), 457–463.
and home treatment for people with mental health problems: a historical cohort Walter, K., Min, J.L., 2015. The UK10K project identifies rare variants in health and
study. Health Soc. Care Community 18 (5), 456–464. disease. Nature 526, 82 (The UK10K Consortium).
Korhonen, K., Remes, H., Martikainen, P., 2017. Education as a social pathway from Watanabe, K., Taskesen, E., 2017. Functional mapping and annotation of genetic
parental socioeconomic position to depression in late adolescence and early associations with FUMA, 8 (1), 1826.
adulthood: a Finnish population-based register study. Soc. Psychiatry Psychiatr. Wendt FR, Pathak GA, Deak JD, De Angelis F, Koller D, Cabrera-Mendoza B, Lebovitch
Epidemiol. 52 (1), 105–116. DS, Levey DF, Stein MB, Kranzler HR, Koenen KC, Gelernter J, Huckins LM,
Polimanti R. Using phenotype risk scores to enhance gene discovery for generalized

10
X. Qi et al. Neuroscience and Biobehavioral Reviews 140 (2022) 104806

anxiety disorder and posttraumatic stress disorder. Mol Psychiatry. 2022 Apr;27(4): Wiszniewski, W., Gawlinski, P., Gambin, T., et al., 2018. Comprehensive genomic
2206-2215. doi: 10.1038/s41380-022-01469-y. Epub 2022 Feb 18. PMID: analysis of patients with disorders of cerebral cortical development, 26 (8),
35181757; PMCID: PMC9133008. 1121–1131.
Werme, J., van der Sluis, S., Posthuma, D., de Leeuw, C.A., 2021. Genome-wide gene- Xia, Y., Ma, Z., 2020. Relative deprivation, social exclusion, and quality of life among
environment interactions in neuroticism: an exploratory study across 25 Chinese internal migrants. Public Health 186, 129–136.
environments. Transl. Psychiatry 11 (1), 180. Zhao, B., Luo, T., Li, T., et al., 2019a. Genome-wide association analysis of 19,629
Westerman, K.E., Lin, J., Sevilla-Gonzalez, Md.R., Tadess, B., Marchek, C., Manning, A. individuals identifies variants influencing regional brain volumes and refines their
K., 2022. Gene-environment interaction analysis incorporating sex, cardiometabolic genetic co-architecture with cognitive and mental health traits, 51 (11), 1637–1644.
diseases, and multiple deprivation index reveals novel genetic associations with Zhao, B., Luo, T., Li, T., et al., 2019b. Genome-wide association analysis of 19,629
COVID-19 severity. Front. Genet. 12. individuals identifies variants influencing regional brain volumes and refines their
Wickham, S., Taylor, P., Shevlin, M., Bentall, R.P., 2014. The impact of social deprivation genetic co-architecture with cognitive and mental health traits. Nat. Genet. 51 (11),
on paranoia, hallucinations, mania and depression: the role of discrimination social 1637–1644.
support, stress and trust. PLOS One 9 (8), e105140.

11