Smoking Initiation Risk and Your Genetics

What is Smoking Initiation Risk?

Smoking initiation risk describes the heritable component of whether a person crosses the threshold from never having smoked regularly to becoming a regular tobacco user. The phenotype studied here uses a binary ever-regular vs never-regular definition: participants who reported ever smoking regularly are classified as cases, while those who never smoked regularly serve as controls. This precise boundary makes the GWAS signal particularly informative about the genetics of the initiation threshold specifically, rather than smoking quantity or dependence severity.

Twin and family studies consistently place heritability for smoking initiation at roughly 50 to 60 percent. The genetic architecture is polygenic — many loci each contributing a small fraction of the heritable component — with enrichment in neurobiological pathways governing reward sensitivity, impulse control, and behavioral reinforcement.

Research base: Robust.

The genetics of Smoking Initiation Risk

Liu et al. (2019), published in Nature Genetics, performed one of the largest GWAS of smoking behaviors, studying up to 1.2 million individuals across multiple cohorts. That landmark analysis covered several smoking phenotypes simultaneously, including the ever-regular vs never-regular binary definition. The scale of the cohort enabled discovery of hundreds of genome-wide significant loci and refined the genetic map of smoking initiation to a degree not previously possible.

Brazel et al. (2019), published in Biological Psychiatry, complemented large-scale common-variant GWAS with an exome chip meta-analysis spanning 152,348 to 433,216 participants across five substance-use phenotypes. By targeting coding and rare variants rather than common intronic signals, that work demonstrated that rare coding variation contributes meaningfully to the heritability of smoking behavior — accounting for roughly 1 to 2 percent of phenotypic variance and approximately 11 to 18 percent of total SNP heritability. Fine-mapping in that study narrowed putative causal variants to single base-pair resolution at 24 loci, providing unusually precise mechanistic anchors in the smoking initiation architecture.

Together, these studies establish smoking initiation as a trait with a rich, multilayered genetic structure — common polygenic variation setting the background level of heritable predisposition, with rarer coding variants adding discrete biological perturbations at specific genes.

Stat block: 1.2 million individuals in the Liu et al. (2019) GWAS of smoking behaviors, including the ever regular vs never regular initiation phenotype.

Stat block: 414 gene-proximal variants captured in the current signal landscape for this ever-regular vs never-regular smoking initiation phenotype.

Key genes: IGSF21, ARID5B, ENTPD6, ALK, and ANKRD26

The gene-level evidence for smoking initiation under the ever-regular vs never-regular definition converges on a set of candidates spanning neural adhesion, transcriptional regulation, purinergic signaling, and receptor tyrosine kinase biology.

IGSF21 (immunoglobulin superfamily member 21) holds the highest confidence ranking for this phenotype. Its strong statistical evidence — including proximity to the lead variant and protein-QTL colocalization — is consistent across multiple smoking initiation phenotypes, making it among the most replicated gene-level findings in tobacco use genetics. IGSF21 encodes a synaptic cell adhesion molecule expressed in the nervous system, where synaptic organization shapes how neural circuits encode reinforcement and reward.

ARID5B (AT-rich interaction domain 5B) is the second-ranked gene by confidence for this phenotype, appearing at loci on chromosome 10 with multiple credible sets. It encodes a chromatin-binding protein involved in transcriptional regulation during development and in response to metabolic and environmental signals. ARID5B appears across multiple tobacco-related GWAS phenotypes, consistent with a regulatory role in gene networks relevant to both initiation and dependence.

ENTPD6 (ectonucleoside triphosphate diphosphohydrolase 6) contributes enzymatic regulation of extracellular purinergic signaling. ATP and its metabolites function as neuromodulators that influence dopaminergic and glutamatergic transmission at synapses in reward-processing regions. ENTPD6's appearance near the top of the gene-confidence ranking for both this phenotype and the MTAG smoking initiation dataset suggests that purinergic signaling modulation is a reproducible biological theme across smoking initiation genetics.

ALK (anaplastic lymphoma kinase) is a receptor tyrosine kinase with a well-characterized role in cancer biology, but its function extends to the nervous system: ALK is expressed in the brain during development and regulates neuronal differentiation and axonal guidance. Emerging evidence links ALK to reward-pathway development, and its appearance in the filtered gene set for smoking initiation is consistent with a developmental neurobiological contribution to addiction susceptibility.

ANKRD26 (ankyrin repeat domain 26) encodes a scaffold protein with roles in signal transduction and cytoskeletal organization. Ankyrin-repeat proteins participate in the assembly of protein complexes at postsynaptic densities, where they influence receptor anchoring and signal integration. Variants near ANKRD26 in the smoking initiation landscape may affect synaptic signaling fidelity in circuits relevant to behavioral reinforcement.

What the research says

The Liu et al. (2019) analysis in Nature Genetics was transformative for tobacco use genetics by virtue of its scale. Studying 1.2 million individuals across smoking initiation, cessation, cigarettes per day, and age of initiation simultaneously revealed genetic correlations among these behaviors and identified loci with effects spanning multiple smoking phenotypes. For the ever-regular vs never-regular phenotype specifically, the genome-wide signal landscape confirmed the deeply polygenic architecture and pointed toward enrichment in neurological and behavioral gene sets.

Brazel et al. (2019) in Biological Psychiatry added a complementary layer by demonstrating that rare coding variation — typically missed by standard SNP arrays — contributes meaningfully to smoking initiation heritability. Fine-mapping at 24 loci to single-variant resolution provided unusually precise mechanistic targets, connecting the polygenic GWAS signal to specific functional coding changes in defined proteins. This convergence of common-variant and rare-variant evidence strengthens confidence in the biological pathways implicated across both analyses.

The consistency of IGSF21 and ARID5B as top-ranked genes across multiple smoking initiation GWAS phenotypes — including this ever-regular vs never-regular definition and broader meta-analyses — provides independent replication of their involvement at the gene level, the strongest form of cross-study validation available in this field.

How Smoking Initiation Risk affects you

A higher genetic score for smoking initiation risk means your genetic profile more closely matches the variant pattern associated with having ever smoked regularly in large population studies. This describes a statistical tendency at the population level — not a prediction of individual behavior.

The biological meaning of a higher score lies in the underlying neuroscience: genetic variation in reward-circuit synaptic organization, transcriptional regulation of neuronal gene expression, and purinergic neuromodulation collectively shape how reinforcing early tobacco exposure is likely to feel. These are upstream biological influences on the probability of initiating regular use, not downstream determinants of a fixed outcome.

Environmental and behavioral factors — social context, household norms, stress, access, and personal choice — remain the dominant proximal determinants of smoking behavior regardless of genetic score.

Working with your Smoking Initiation Risk profile

• If you currently smoke, this genetic profile provides biological context. All major evidence-based cessation strategies — pharmacological and behavioral — are effective across the genetic spectrum.
• For those who have never smoked regularly, a higher genetic score is informative about biological predisposition but confers no certainty about future behavior. Social and situational factors remain the primary modifiable influences.
• Share your genetic profile with a clinician if you are in a cessation program or have a strong family history of tobacco dependence — it can complement a holistic clinical assessment.
• Because IGSF21 and ARID5B signals appear across multiple smoking phenotypes, a higher score here may be worth considering alongside related trait scores for a fuller picture of tobacco use genetics.

Frequently asked questions

Q: What does 'ever regular vs never regular' mean as a phenotype? A: It is a binary classification: people who self-reported ever smoking on a regular basis are cases, and those who never smoked regularly are controls. This clean binary captures the threshold-crossing moment of initiation — which is genetically distinct from how much someone smokes or whether they can quit.

Q: Why do IGSF21 and ARID5B appear across multiple smoking initiation studies? A: Both genes sit near variants that reach genome-wide significance in multiple independent smoking initiation GWAS with different phenotype definitions and ancestry compositions. Replication across independent studies is the strongest evidence available in this field that the gene-level association is genuine rather than a statistical artifact.

Q: What role does ALK play in smoking initiation genetics? A: ALK is best known in oncology, but in the nervous system it is a receptor tyrosine kinase involved in neuronal differentiation and axonal guidance during brain development. Emerging evidence links ALK to reward-pathway formation, and its presence in the smoking initiation gene set points to a developmental neurobiological contribution to addiction susceptibility.

Q: Does this genetic score apply only to cigarette smoking? A: The GWAS studies underlying this trait defined smoking initiation as ever using tobacco cigarettes regularly. The genetic signals reflect biology relevant to nicotine and reward-circuit development broadly, but the phenotype definition was cigarette-specific. Whether the score generalizes to other tobacco products has not been directly tested in the studies cited here.

Q: Is the research quality for this trait strong? A: Yes — this trait carries a robust confidence tier. The Liu et al. (2019) analysis studied 1.2 million individuals, providing exceptional statistical power, and Brazel et al. (2019) independently confirmed contributions from rare coding variants. The convergence of common- and rare-variant evidence across large independent cohorts meets a high bar for confidence.

References

Liu M, et al. (2019). Association studies of up to 1.2 million individuals yield new insights into the genetic etiology of tobacco and alcohol use. Nat Genet. PMID: 30643251. Brazel DM, et al. (2019). Exome Chip Meta-analysis Fine Maps Causal Variants and Elucidates the Genetic Architecture of Rare Coding Variants in Smoking and Alcohol Use. Biol Psychiatry. PMID: 30679032.

Data sources: GWAS Catalog, Open Targets, ClinVar, ClinGen, NCBI Gene, dbSNP, PheGenI.

Additional research: smoking behavior

The genetics behind smoking behavior

Smoking behavior tendency is highly polygenic — hundreds of genetic variants distributed across the genome collectively shape the biological susceptibility to tobacco initiation. No single gene determines whether a person smokes; rather, an accumulating profile of small genetic effects shapes the biological landscape in which tobacco exposure either takes hold or doesn't.

Among the stronger genetic signals in this analysis is a region near ARID5B (AT-rich interaction domain 5B), a transcription factor expressed in brain tissues and immune cells. ARID5B regulates the expression of other genes involved in neural development and stress response — functions that may link it to the behavioral systems underlying risk-taking and addiction initiation.^[²]

Also in the associated gene set is ALDH1B1 (aldehyde dehydrogenase 1B1), an enzyme that metabolizes aldehydes — a category of reactive compounds that includes those produced during tobacco combustion. Genetic variation in ALDH1B1 may influence how the body processes tobacco-derived aldehydes, potentially shaping the subjective experience of early smoking and how aversive or reinforcing that experience is.

ADGRB2 (adhesion G protein-coupled receptor B2) is another gene in the association set. ADGRB2 is expressed specifically in the brain, where it plays a role in angiogenesis regulation and neuronal signaling. Its brain-specific expression places it in the neurological systems that tobacco use research consistently implicates in initiation and dependence.

Multiple independent chromosomal regions associated with smoking initiation have been identified in genome-wide analyses spanning hundreds of thousands of individuals — confirming the highly polygenic architecture of tobacco use behavior and the real but modest contribution of inherited factors to smoking onset.^[¹][²]

Population-level genetic factors influencing smoking initiation appear to act through convergent pathways: the mesolimbic dopamine system (which determines how reinforcing first nicotine exposures are), stress-response circuitry (which shapes tobacco's appeal as a coping mechanism), and neurological factors related to impulsivity and novelty-seeking. Variants across all of these systems contribute small effects that accumulate into an overall genetic tendency.^[¹]

What the research says

Research base: Robust. The genetics of smoking initiation is among the most extensively studied behavioral-genetic questions, with replication across multiple large cohorts, diverse ancestral backgrounds, and independent research groups. The "ever vs never smoker" phenotype has clean epidemiological definition, enabling population-scale comparisons.

Large-scale genome-wide analyses — including studies involving hundreds of thousands of participants — have identified robust genetic associations with smoking initiation in loci distributed across the genome.^[¹][²] Polygenic scores constructed from these associations predict statistically meaningful differences in smoking rates across population quintiles, though with substantial individual variation within each quintile.

Heritability estimates for smoking initiation range from approximately 40 to 60 percent in twin and family studies — placing it firmly among moderately heritable behavioral traits, alongside personality dimensions and other substance use phenotypes. Environmental factors — access to tobacco, social environment, early exposures, marketing — account for the remaining variation.^[³]

For the statistical methods behind variant selection and polygenic scoring — see our methodology page for the full approach.

References (smoking behavior)

1. Clifton EAD, et al. (2018). Genome-wide association study for risk taking propensity indicates shared pathways with body mass index. Commun Biol. PMID: 30271922.
2. Karlsson Linnér R, et al. (2019). Genome-wide association analyses of risk tolerance and risky behaviors in over 1 million individuals identify hundreds of loci and shared genetic influences. Nat Genet. PMID: 30643258.
3. Cai N, et al. (2020). Minimal phenotyping yields genome-wide association signals of low specificity for major depression. Nat Genet. PMID: 32231276.

Data sources:

• GWAS Catalog (NHGRI-EBI, accessed 2026-05-20)
• Open Targets Platform (CC0 1.0, accessed 2026-05-20)
• ClinVar (NCBI, accessed 2026-05-20) — entries at ≥2-star review status
• ClinGen Gene-Disease Validity (CC0 1.0, accessed 2026-05-20)

Additional research: smoking likelihood

The genetics of Smoking Likelihood

Research base: Robust.

Xu et al. (2023), published in Addiction, applied MTAG to genetic data from 49,929 individuals drawn from the Yale-Penn cohort and the Penn Medicine BioBank, spanning European and African ancestry participants. The analysis jointly modeled smoking initiation alongside opioid use disorder, cannabis use disorder, and alcohol use disorder — phenotypes that share substantial genetic architecture with tobacco initiation. By leveraging the genetic correlations among these behaviors, MTAG amplified statistical power beyond what the sample size alone would deliver.

The cross-phenotype framework identified genome-wide significant loci near ENTPD6, LMO3, and XYLT1 that emerged with higher confidence in the multi-trait analysis than in smoking-only analyses of comparable size. Polygenic risk scores derived from MTAG showed stronger predictive power for each individual substance-use trait than scores from single-trait GWAS, demonstrating the practical value of cross-phenotype enrichment for discovery and for score construction.

Genes identified through MTAG for smoking initiation carry a distinct interpretive nuance: their signals are enriched because multiple addiction-related phenotypes converge on overlapping loci, suggesting involvement in shared neurobiological pathways rather than smoking-specific mechanisms. This makes them informative for understanding addiction vulnerability broadly, not only tobacco use.

Stat block: 49,929 individuals in the Xu et al. (2023) multi-trait GWAS contributed to identifying shared genetic architecture across substance-use behaviors including smoking initiation.

Stat block: 182 gene-proximal variants in the refined MTAG signal set for this smoking initiation phenotype, reflecting the higher specificity of the cross-phenotype approach compared to broader single-trait catalogs.

What the research says

The Xu et al. (2023) MTAG analysis in Addiction was designed to address a fundamental challenge in addiction genetics: the substantial genetic overlap among substance-use disorders. By treating this overlap as a resource rather than a confound, MTAG simultaneously improves discovery power for each individual phenotype and reveals the genes most likely to operate through shared mechanisms.

For smoking initiation specifically, the multi-trait analysis recovered loci that would have fallen below genome-wide significance in a single-trait GWAS of comparable sample size. This is particularly valuable for behavior genetics, where sample sizes tend to be smaller than for physiological traits. The enrichment of loci near neurodevelopmental genes — AUTS2, ARID5B, ATXN1L — is consistent with the hypothesis that early-life neural circuit formation sets a biological baseline for addiction susceptibility that persists into adulthood.

The cross-ancestry design of the Xu et al. cohort adds further confidence to loci that replicate across European and African ancestry participants. Variants surviving cross-ancestry analysis are less likely to reflect population-stratification artifacts and more likely to represent genuine functional effects at the implicated genes.

The compact gene set from this MTAG analysis — 182 gene-proximal variants compared to over 1,900 in broader smoking initiation meta-analyses — reflects the precision gain from the multi-trait approach. A smaller, higher-confidence gene list is more tractable for biological follow-up and for understanding which pathways are most centrally involved in the shared genetic risk for substance use behaviors.

References (smoking likelihood)

1. Xu H, et al. (2023). Identifying genetic loci and phenomic associations of substance use traits: a multi-trait analysis of GWAS (MTAG) study. Addiction. PMID: 37156939.

Additional research: smoking status

The genetics behind Smoking Status

The genetic architecture of smoking status involves multiple neurobiological pathways, with the strongest signals concentrated in genes related to dopamine reward processing, epigenetic regulation, synaptic connectivity, and appetite and stress biology.

RASGRF2 (Ras protein-specific guanine nucleotide-releasing factor 2) is the top-ranked genetic signal for smoking status in this analysis. RASGRF2 encodes a guanine nucleotide exchange factor that activates Ras-MAPK signaling in neurons — pathways critical for dopamine receptor signaling and synaptic plasticity in reward circuits. RASGRF2 has appeared in genome-wide analyses of multiple substance use behaviors and alcohol use, consistent with a role in the general neurobiological substrates of reward-seeking behavior. Its appearance as the top signal for smoking status underscores the centrality of dopamine reward biology in tobacco use initiation. (Kichaev et al. 2019)[1]

JADE2 (jade family PHD finger 2) encodes a chromatin-associated protein containing a PHD finger domain that is involved in histone acetylation and epigenetic gene regulation. Epigenetic regulation of gene expression in neurons — particularly through histone modifications — shapes the long-term transcriptional states of reward circuit neurons. JADE2's appearance in smoking status genetics points to epigenetic mechanisms as contributors to smoking initiation predisposition, potentially through regulation of gene expression in dopaminergic or other reward-relevant neuronal populations. (Kichaev et al. 2019)[1]

BARHL2 (BarH-like homeobox 2) is a homeobox transcription factor required for the development and specification of inhibitory neuronal populations. BarH-like factors are important for the identity and connectivity of neurons in brain regions relevant to reward processing and behavioral regulation. Variants near BARHL2 appear across multiple genome-wide smoking analyses — both lifetime exposure and smoking status — suggesting it represents a robust signal in the genetic architecture of tobacco use behavior. (Kichaev et al. 2019)[1]

NUCB2 (nucleobindin 2) encodes the precursor protein for nesfatin-1, a peptide involved in regulating appetite and stress-related feeding behavior. Nesfatin-1 acts in the hypothalamus and brainstem to influence energy balance and emotional regulation, and has connections to dopaminergic reward signaling. NUCB2 variants in the context of smoking status genetics point to an intersection between stress regulation, appetite biology, and the neurobiological circuits underlying tobacco use behavior. (Kichaev et al. 2019)[1]

NYAP2 (neuronal tyrosine-phosphorylated phosphoinositide-3-kinase adaptor 2) encodes an adaptor protein that activates PI3K/Akt signaling specifically in neurons. PI3K/Akt signaling regulates neuronal survival, axon growth, and synaptic plasticity. NYAP2's appearance in smoking status genetics reflects the contribution of neuronal growth and plasticity pathways to the development of circuits underlying smoking behavior. (Kichaev et al. 2019)[1]

BDNF (brain-derived neurotrophic factor) appears in the gene set for smoking status, consistent with its replicated role in smoking and addiction genetics broadly. BDNF supports the survival and differentiation of dopaminergic neurons and regulates synaptic plasticity in reward circuits. Its presence across multiple smoking-related phenotypes — including smoking status — reflects its foundational role in the neurobiology of reward and behavioral persistence. (Kichaev et al. 2019)[1]

Genome-wide analyses of smoking status identify signals near RASGRF2, JADE2, BARHL2, NUCB2, and NYAP2 — reflecting dopamine reward signaling, epigenetic regulation, neurodevelopment, and stress-appetite biology as contributors to the neurobiological basis of tobacco use initiation. (Kichaev et al. 2019)^[1]

What the research says

Research base: Moderate. Smoking status genetics is supported by genome-wide evidence from multiple studies and ancestries. The moderate confidence tier reflects that smoking behavior — including initiation — is influenced by powerful social and environmental factors alongside genetics, and that individual genetic signals have small effects consistent with a polygenic architecture.

Kichaev et al. (2019) applied a functional enrichment approach (FINDOR) to genome-wide association analyses across 27 complex traits, including smoking status from UK Biobank data. Their method leverages functional annotations to improve statistical power for identifying genetic signals in large biobank-scale datasets. Smoking status was among the traits showing meaningful improvement in detected loci using this approach, contributing to the catalog of genetic signals for tobacco use behavior. (Kichaev et al. 2019)[1]

The broader smoking genetics literature — including studies of smoking initiation, cigarettes per day, and nicotine dependence — consistently identifies dopamine reward pathway genes, GABAergic interneuron development genes, and synaptic biology genes as the core architecture of tobacco use genetics.

Twin studies of smoking initiation estimate heritability at 40–60 percent. Genome-wide research consistently identifies dopamine signaling, GABAergic circuitry, and synaptic plasticity as the primary neurobiological substrates of inherited tobacco use predisposition. (Kichaev et al. 2019)^[1]

References (smoking status)

1. Kichaev G, Bhatia G, Loh PR, et al. (2019). Leveraging polygenic functional enrichment to improve GWAS power. American Journal of Human Genetics. PMID: 30595370.