INTRODUCTION
Although sleep represents one of the most fundamental biological processes, individual differences in sleep patterns, quality, and timing remain poorly understood. While environmental factors significantly influence sleep, mounting evidence suggests that genetic factors also contribute substantially to individual variations in sleep phenotypes. Twin studies have demonstrated that 46% of the variability in sleep duration and 44% of the variability in sleep quality are genetically determined [1,2].
Genetic architecture of sleep regulation
The molecular basis of sleep regulation involves intricate networks of genes that control circadian rhythms, neurotransmitter systems, and metabolic pathways. Core circadian clock genes, including CLOCK, BMAL1, PER1-3, and CRY1-2, establish the fundamental 24 h biological rhythm through negative feedback loops [3,4]. Recent genome-wide association studies (GWAS) have identified 351 genetic loci associated with chronotype preferences, many of which are enriched for circadian regulation and neurotransmitter signaling pathways [5].
Beyond traditional circadian control, genetic variants affecting metabolic pathways may significantly affect sleep regulation. Calcium homeostasis plays a crucial role in neuronal excitability and neurotransmitter release, with studies showing that suprachiasmatic nucleus neurons exhibit circadian rhythms in intracellular calcium concentration [6]. In addition, coenzyme Q10 (CoQ10) serves essential functions in mitochondrial energy production [7] and is known to help improve daytime fatigue, one of the daytime symptoms of sleep apnea and other sleep disorders [8,9]. Genetically determined levels of nicotine metabolism can influence an individual’s nicotine consumption, which can lead to sleep problems such as reduced total sleep time and daytime sleepiness [10,11]. In this way, genetic variations related to nutrient and material metabolism can play a decisive role in sleep phenotypes by influencing how individuals interact with their environment, such as through lifestyle habits. Therefore, it is necessary to go beyond simply exploring genetic markers that influence the manifestation of sleep disorders and instead investigate the relationships between broader genetic markers that influence an individual’s lifestyle and sleep patterns. Such an exploration is important because it may provide key clues to explaining individual differences in sleep patterns.
Such exploratory studies may provide new insights into the relationship between disease susceptibility and sleep quality. Although many previous studies have demonstrated the impact of sleep on fatal diseases such as cancer [12], these relationships have not shown consistent results [13,14]. A comprehensive exploration of the relationship between genetic markers associated with health and disease in healthy adults and sleep traits could provide new clues to understanding the relationship between disease susceptibility and sleep, which has not yet been fully explained by a one-to-one relationship.
Study rationale and objectives
Most genetic studies have focused on single candidate genes or pathological sleep conditions, leaving gaps in the understanding of how genetic variants influence sleep patterns in healthy populations. The present study addressed this issue by examining the relationship between comprehensive genetic risk profiles and sleep parameters in healthy young Korean adults. We employed a broad approach to examine lifestylerelated genetic markers (calcium homeostasis, CoQ10 levels, chronotype, and nicotine metabolism) and disease-related genetic markers (cancer susceptibility variants).
Our approach combines multiple analytical methods— Spearman correlations, group-based analysis of variance (ANOVA), and individual single-nucleotide polymorphism (SNP) analyses—to ensure robust findings across different levels of genetic analysis. This study aimed to 1) characterize the relationships between lifestyle-related genetic markers and sleep parameters, 2) examine the associations between disease susceptibility variants and sleep outcomes, 3) identify specific genetic variants with the strongest effects on sleep phenotypes, and 4) explore the potential mechanisms underlying the observed genetic-sleep relationships.
Understanding these genetic influences has significant implications for personalized sleep medicine. If genetic variants can reliably predict individual differences in sleep patterns and treatment responses, this information could guide more effective individualized interventions. This research represents a crucial step toward integrating genetic information into clinical sleep medicine, moving beyond one-size-fits-all approaches toward personalized treatment strategies that optimize outcomes for each patient’s unique biological profile.
METHODS
Study design and procedures
This investigator-initiated, single-center, cross-sectional observational study was conducted at the International Saint Mary’s Hospital in Incheon, Republic of Korea, from December 2023 to December 2024. The study protocol was approved by the Institutional Review Board of the International Saint Mary’s Hospital (IRB No. IS23TSSE0080), and written informed consent was obtained from all participants.
Eligible participants who met the inclusion criteria were assigned a unique identification code and underwent anonymization. Demographic data (age, sex, educational level, and occupation) were collected. Participants visited the hospital to complete baseline questionnaires regarding sleep quality and underwent direct-to-consumer genetic testing on the same day. Subsequently, they completed a 2-week online sleep diary at home.
Participants and recruitment
The study population consisted of healthy adults aged 20– 39 years in early adulthood with no history of serious medical or psychiatric conditions. The participants were recruited through hospital bulletin boards and YouTube channels dedicated to sleep-related content. Those who voluntarily expressed interest in participating were informed about the study and provided written informed consent for the use of human-derived materials and genetic testing before undergoing screening.
Participants were excluded if they met any of the following criteria during screening: 1) current or past treatment for medical conditions (e.g., cardiovascular, renal, neurological, thyroid, gastrointestinal diseases, diabetes, hypertension, cancer) or psychiatric conditions (e.g., depression, bipolar disorder, panic disorder, anxiety disorders, or psychosis); 2) use of sleeprelated medications or a visit to a sleep clinic within the past year or a history of suicide attempts or self-harm; 3) shift work; or 4) pregnancy or breastfeeding.
Self-report questionnaires and sleep diary
Insomnia Severity Index
The Insomnia Severity Index (ISI) was developed by Bastien et al. [15] and consists of 7 items that measure the severity of insomnia during the past 2 weeks. The ISI is rated on a 5-point Likert scale ranging from 0–4. The total score ranges from 0–28, with higher scores indicating greater insomnia severity.
Pittsburgh Sleep Quality Index
The Pittsburgh Sleep Quality Index (PSQI) was developed by Buysse et al. [16] and is a self-report assessment tool that evaluates sleep quality over a 1-month period. A global score and seven component scores can be derived from this scale. The component scores include the following: subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleeping medications, and daytime dysfunction. Each component is scored on a scale of 0–3, with a total score ranging from 0–21, where a higher score indicates poorer sleep quality.
Morningness–Eveningness Questionnaire
The Morningness–Eveningness Questionnaire (MEQ) is the most commonly used circadian typology questionnaire and consists of 19 questions with Likert responses about an individual’s preferred bedtimes, get-up times, and times for activity [17]. Responses to the questions are scored and summed to produce an overall morningness score ranging from 16–86, with higher scores indicating greater morningness [18].
Epworth Sleepiness Scale
The Epworth Sleepiness Scale (ESS) consists of 8 items that measure excessive daytime sleepiness. This scale was developed by Johns (1991) [19] and the items are rated on a 4-point Likert scale ranging from 0–3. The total score ranges from 0–24 points, with higher scores reflecting greater levels of daytime sleepiness. In this study, Cronbach’s α for the ESS was 0.72.
Maslach Burnout Inventory
The Maslach Burnout Inventory (MBI) measures job burnout defined by three subscales: emotional exhaustion (EE, 9 items), depersonalization (DP, 5 items), and professional accomplishment (PA, 8 items), each rated on a 7-point Likert scale ranging from 0–6 [20,21]. Higher scores on the EE and DP subscales indicate a greater burnout symptom burden; lower scores on the PA subscale indicate a higher burnout symptom burden.
Sleep diary
Participants completed daily sleep diaries for 14 consecutive days, recording their bedtime (BT), lights-out time (LO), sleep onset latency in minutes (SOL), wake after sleep onset in minutes (WASO), final wake time (WT), time out of bed (TOB), time in bed in minutes (TIB), total sleep time in minutes (TST), sleep efficiency (SE, calculated as TST/TIB×100%), mid-point of sleep (MP), subjective sleep quality (SQ), and morning freshness (FRESH). The average values were computed for each participant across all sleep parameters, and no outlier removal was conducted to preserve the natural variance within the healthy population.
Genetic testing: using direct-to-consumer genetic testing panels
Buccal swab samples were collected using the Buccal DNA Collector (Theragen Bio Co., Ltd.). The Buccal DNA Collector collects DNA samples from participants [22] and is as accurate as blood sample testing [23,24]. Therefore, in the present study, the collector was used to perform accurate genetic testing while minimizing the risk to participants. Genomic DNA was extracted using a commercial extraction kit (Gene All Biotechnology Co., Ltd.) according to the manufacturer’s instructions. The concentration and purity of the extracted DNA were measured using a Tecan F200 microplate reader (Tecan Austria GmbH). The microarray experiments were conducted at the Microarray Core Facility of Theragen Health Co., Ltd. (Seongnam, Korea).
Whole-genome amplification was performed on the extracted DNA, followed by random fragmentation into 25–125 bp segments. The fragmented DNA was purified, resuspended, and hybridized to the Theragen Precision Medicine Research Array version 3 (Theragen PMRA3), a customized microarray based on the Affymetrix platform (Thermo Fisher Scientific).
After hybridization, nonspecifically bound targets were removed under stringent washing conditions to minimize background signals arising from random ligation events. Genotyping of approximately 150,000 SNPs was performed using the PMRA3 array according to the manufacturer’s protocol. The array provides genome-wide coverage across five major global populations and supports high-accuracy genotype imputation, achieving imputation accuracies of 0.90 and 0.94 for minor allele frequencies >1% and >5%, respectively, based on a reference panel of 7.4 million imputed markers in the Asian population.
The sample quality control (QC) criteria for genotyping were as follows: 1) DNA QC ≥0.82, 2) QC call rate ≥97%, and 3) average call rate for passing samples ≥98.5%. Low-quality SNP genotyping data were excluded based on the QC cutoff thresholds, including 1) cr-cutoff <95%; 2) fld-cutoff <3.6; 3) het-so-cutoff <-0.1; 4) hom-ro-1-cutoff <0.6; 5) hom-ro-2-cutoff <0.3; or 6) hom-ro-3-cutoff <-0.9.
Statistical analysis
All statistical analyses were conducted using IBM SPSS Statistics (version 21.0; IBM Corp.). Descriptive statistics (mean± standard deviation [SD] or frequency and percentage) were computed for demographic and sleep-related variables. A total of 85 samples were included in the final analysis, excluding one subject who took melatonin, a drug that affects sleep, during the study period.
To examine the relationship between genetic markers and sleep parameters, three complementary analytical approaches were applied. First, Spearman’s rank correlation coefficients were calculated between ordinal genetic risk scores (categorized as Good–Caution–Warning or Good–Caution–Warning– Intensive) and sleep parameters. Risk classifications were based on Theragen Health’s proprietary system, which integrates GWAS and functional annotation data. To address multiple testing across approximately 600 genetic markers, we adopted a conservative approach rather than a formal false discovery rate (FDR) correction. Only associations that showed statistical significance (p<0.05) and consistent directional trends across ≥3 sleep parameters were retained for the main interpretation. For the directional consistency criterion, correlations were required to show the same direction (positive or negative) across multiple sleep parameters within each genetic marker. For disease-related markers, skin cancer susceptibility genes (basal cell carcinoma, malignant melanoma, and squamous cell carcinoma) were evaluated as a single phenotypic category due to their shared pathophysiological mechanisms and consistent association with superior sleep quality and morning chronotype preference. Second, one-way ANOVA was conducted to compare sleep variables across the genetic risk groups. Levene’s test was used to assess the homogeneity of variance. When variances were equal (p>0.05), Bonferroni post-hoc tests were used; when unequal (p≤0.05), Games–Howell tests were applied. All comparisons included effect sizes and 95% confidence intervals (CIs).
Finally, genotype-wise ANOVAs were performed (e.g., AA vs. AG vs. GG) to select SNPs flagged by prior reports or strong group-level results. Post-hoc analyses followed the same method described above. SNPs were prioritized based on their presence in Theragen’s health panels and their previously reported associations with sleep or neurophysiology.
Missing data and multiple testing
Participants with missing genotypes or diary data were excluded during preprocessing. Thus, the final analysis included no missing data. To address multiple comparisons, we employed a dual strategy: 1) effect size-based prioritization, where associations with stronger correlation coefficients were given greater interpretive weight and represented with higher color intensity in visualizations, and 2) replication of effects across multiple sleep parameters with consistent directional trends (≥3 parameters).
RESULTS
Demographic characteristics and sleep patterns
The final sample comprised 85 participants with a mean age of 28.18 years (SD=4.3) who were predominantly female (77.6%), highly educated (80.0% college-educated), and mostly employed (67.1%). Despite this healthy population, sleep quality patterns were concerning: PSQI scores averaged 6.0 (SD=2.2), exceeding the clinical cutoff of 5. The sleep diary data revealed a generally adequate sleep architecture, with an average SOL of 14.1 min (SD=8.1), WASO of 8.3 min (SD=10.7), and SE of 94.7% (SD=3.7%). Most participants (89.4%) displayed intermediate chronotype preferences, with no identified true morning types. The demographic characteristics are shown in Table 1.
Correlations between genetic markers and sleep parameters
As a result of Spearman correlations, significant and consistent associations with three or more sleep parameters were found for the following genetic markers: calcium deficiency, CoQ10 deficiency, nicotine metabolism, chronotype, gallbladder cancer, and skin cancer. The results are shown in Fig. 1.
Among the health-related genetic markers, calcium deficiency was significantly and positively associated with BT (r=0.241, p<0.05), LO (r=0.248, p<0.05), and WT (r=0.225, p< 0.05). CoQ10 deficiency was significantly associated with TIB (r=0.244, p<0.05), TST (r=0.228, p<0.05), TOB (r=0.236, p<0.05), WT (r=0.259, p<0.05), and PSQI scores (r=0.227, p< 0.05). Nicotine metabolism had a significant positive association with SOL (r=0.229, p<0.05), WASO (r=0.356, p<0.01), and a significant negative association with SE (r=-0.343, p<0.01), FRESH (r=-0.261, p<0.05), and PSQI (r=0.291, p<0.01). The chronotype was also significantly associated with BT (r=0.232, p<0.05), WT (r=0.230, p<0.05), and TOB (r=0.240, p<0.05).
Correlation analyses revealed strong associations between disease-related genetic markers and sleep parameters. Gallbladder cancer showed a significant positive association with SOL (r=0.341, p<0.01) and PSQI (r=0.245, p<0.05) and significant negative association with SE (r=-0.335, p<0.01). Basal cell carcinoma had a significant negative association with LO (r=-0.219, p<0.05), WT (r=-0.249, p<0.05), TOB (r=-0.265, p<0.05), and bedtime procrastination (BIP) (r=-0.234, p<0.05). Furthermore, basal cell carcinoma also had a significant positive association with MEQ (r=0.214, p<0.05), and demonstrated a negative association with PSQI (r=-0.249, p<0.05) and MBI_E (r=-0.249, p<0.05). Malignant melanoma was significantly associated with PSQI (r=-0.245, p<0.05), MEQ (r=0.246, p<0.05), SE (r=0.223, p<0.05), FRESH (r=0.271, p<0.05), and SQ (r=0.374, p<0.001). Squamous cell carcinoma was also significantly associated with TIB (r=0.236, p<0.05), TST (r=0.231, p<0.05), and ESS (r=-0.215, p<0.05).
Comparing sleep parameters in genetic markers
Results of the comparison of sleep variables across genetic risk groups are demonstrated in Table 2 and Fig. 2.
There was a significant difference in MP among the three genetic risk groups (good: n=22; caution: n=46; warning: n=17) for calcium deficiency (F=4.341, p=0.020). The Games-Howell post hoc test showed that the “caution” group required significantly more time to get out of bed than the “good” group (mean difference [MD]=6.223 min, p=0.020).
In addition, a one-way ANOVA revealed significant group differences in WT (F=4.750, p=0.011) and ESS (F=12.800, p=0.014) among the three genetic risk groups with CoQ10 deficiency (good: n=66; caution: n=17; warning: n=2). The post hoc test showed that the “caution” group awakened 48 min later than the “good” group (Bonferroni test, p=0.044), and the “warning” group exhibited significantly higher ESS scores than the “caution” group (Games-Howell test, MD=3.560, p=0.013).
A comparison of the three genetic risk groups for nicotine metabolism (fast: n=22; medium: n=46; slow: n=17) demonstrated statistically significant differences in FRESH (F=5.503, p=0.006), WASO (F=4.709, p=0.016), SE (F=5.553, p=0.005), and PSQI score (F=4.518, p=0.014). Specifically, the Bonferroni post hoc test showed that the “fast” metabolism group achieved significantly better morning refreshment than both the “medium” (MD=0.577, p=0.018) and “slow” (MD=0.686, p=0.005) groups. The “slow” group showed substantially lower SE than both the “medium” (Bonferroni test, MD=2.256%, p=0.020) and “fast” (Bonferroni test, MD=3.318%, p=0.022) groups. The “slow” group showed substantially lower sleep quality than the “fast” group (Bonferroni test, MD=2.121, p=0.015). In addition, the Games-Howell post hoc test also showed that the “slow” metabolism group had significantly higher WASO than both the “fast” (MD=8.959 min, p=0.041) and “medium” (MD=6.924 min, p=0.014) groups.
Unexpected results were observed when comparing the genetically determined chronotype groups (morning type: n=3; intermediate type: n=28; evening type: n=54). SOL showed significant group effects (F=4.044, p=0.021), with the genetically determined “morning” group demonstrating substantial-ly longer sleep onset than both the “intermediate” (MD=12.950 min, p=0.024) and “evening” (MD=13.160 min, p=0.017) groups. Even more strikingly, SE showed pronounced group differences (F=6.263, p=0.003), with the “morning” group achieving notably lower efficiency than both the “intermediate” (MD=6.860%, p=0.005) and “evening” (MD=7.238%, p=0.002) groups.
Moreover, significant group differences were found in disease-related genetic marker groups, such as gallbladder cancer and three types of skin cancer. Comparison of the four genetic risk groups for gallbladder cancer (good: n=37; caution: n=32; warning: n=14; intensive: n=2) demonstrated statistically significant differences in SOL (F=10.095, p=0.004) and SE (F=7.815, p=0.002). The Games-Howell post hoc test found that the “good” genetic risk group fell asleep significantly faster than both the “caution” (MD=5.666 min, p=0.017) and “intensive” (MD=9.034 min, p=0.046) groups, and achieving substantially higher efficiency than the “intensive” group (MD=2.952%, p<0.001).
Comparison of the four genetic risk groups for malignant melanoma (good: n=54; caution: n=25; warning: n=6; intensive: n=0) demonstrated significant group differences in SQ (F=7.540, p=0.001). Results of the Bonferroni post hoc test showed that the “good” genetic risk group demonstrated superior sleep quality compared to the “warning” group (MD=0.799 points, p=0.002). Although a one-way ANOVA was planned to compare variables across risk groups of basal cell carcinoma and squamous cell carcinoma, the analysis could not be conducted due to extremely imbalanced group distributions.
Individual SNP analysis results for sleep parameters
Results of the individual SNP analyses are demonstrated in Table 3.
Individual SNP analyses revealed additional complexity in the calcium-sleep relationship. The rs1570669 variant (CYP24A1 gene, affected allele “A”) showed a significant effect on ISI scores (F=3.631, p=0.031), with AG (n=43) carriers exhibiting higher ISI scores than GG (n=25) carriers. More intriguingly, rs34339006 (DGKD gene, affected allele “T”) demonstrated a significant effect on daytime sleepiness (F=5.199, p=0.007), but in an unexpected direction: TT (n=4) carriers (highest genetic risk) had significantly lower ESS scores compared to both TC (n=33; MD=4.530, p=0.006) and CC (n=48; MD=3.771, p=0.025) carriers. The rs780094 variant (GCKR gene, affected allele “C”) showed significant associations with perceived burnout (F=4.726, p=0.011), with CC carriers reporting higher burnout scores than TT carriers (MD=4.170, p=0.009).
Individual SNP analysis of rs933585 (NRXN1 gene, affected allele “A”) related to CoQ10 provided molecular-level confirmation of these group patterns. The variant showed a significant effect on daytime sleepiness (F=12.80, p=0.014) and waking timing (F=4.75, p=0.011). AA carriers had significantly higher ESS scores than AG carriers (MD=3.56, p=0.013), whereas AG carriers exhibited significantly longer WT than GG carriers (MD=0:48, p=0.044). This heterozygous pattern suggests that moderate genetic risk (AG) primarily affects timing, whereas high risk (AA) impairs daytime alertness.
Individual SNP analysis of rs56113850 (CYP2A6 gene, affected allele “C”) related to nicotine metabolism confirmed these group findings, with CC carriers (fastest metabolism) scoring higher than both TT (MD=0.795, p=0.011) and TC carriers (MD=0.740, p=0.020).
Individual SNP analyses have revealed additional complexities in chronotype genetics. The rs11708779 variant (ERC2 gene, affected allele “G”=eveningness) showed significant effects on TIB (F=4.415, p=0.015), with GG carriers spending less TIB than GA carriers (MD=23.960 min, p=0.016). Conversely, rs10493596 (AK5 gene, affected allele “C”=eveningness) demonstrated opposite effects on TST (F=5.034, p=0.009), with CC carriers achieving longer sleep duration than CT carriers (MD=25.705 min, p=0.032). The findings regarding health-related genetic markers and sleep variables are shown in Fig. 3.
Individual SNP analysis of rs1052133 (OGG1 gene, affected allele “G”) related to gallbladder cancer provided molecu-lar-level validation of these group patterns. The variant showed significant effects on SOL (F=7.185, p=0.002), with CC carriers (lowest genetic risk) having a substantially shorter sleep onset than both GG (MD=6.971 min, p=0.007) and GC (MD=4.826 min, p=0.023) carriers.
Individual SNP analysis of rs7023329 (MTAP gene, affected allele “A”) related to malignant melanoma provided molecular validation, with AA carriers (highest melanoma risk) achieving significantly higher MEQ scores than GG carriers (MD=3.080 points, p=0.046). The findings regarding diseaserelated genetic markers and sleep variables are shown in Fig. 4.
DISCUSSION
Integration of findings
The comprehensive analysis revealed three distinct genetic-sleep phenotypes: 1) compensatory patterns in which metabolic deficiencies (calcium and CoQ10) were associated with altered sleep architecture; 2) pharmacological validation of nicotine metabolism; and 3) disease susceptibility paradoxes, where variants associated with health risks demonstrated superior sleep outcomes.
These discoveries have profound implications for personalized medicine approaches and our understanding of how evolution has shaped human physiological systems to balance competing selective pressures. The identification of genetic variants that simultaneously predict disease susceptibility and sleep advantages suggests that the genetic architecture involves complex trade-offs between different physiological systems, where optimization for one domain may inadvertently increase vulnerability in another.
Key findings and mechanisms
Our study revealed paradoxical relationships between genetic markers and sleep outcomes, challenging the conventional understanding of health-related genetics. The most striking finding was that a genetic predisposition to the morning chronotype was associated with poor sleep quality, con-tradicting decades of research linking morning preference to superior sleep health [25-28].
This discrepancy could be attributed to several factors. First, this cross-sectional study of healthy adults without clinical symptoms may not have captured pronounced differences in sleep and psychological indicators between the chronotype groups. In clinical populations or individuals with sleep disorders, chronotype differences are typically more pronounced, owing to the amplification of the underlying circadian misalignment by pathological conditions. However, our study population consisted of relatively healthy individuals with normal sleep patterns (averaging 6–7 h of sleep), which may have created a “ceiling effect” in which the beneficial aspects of the morning chronotype were masked by the generally adequate sleep health of all participants. Additionally, healthy populations often demonstrate greater behavioral flexibility and compensatory mechanisms that can mitigate genetic predispositions, potentially obscuring the true impact of chronotype genetics on sleep outcomes. Second, the study design reflected sleep status at a specific time point, potentially failing to capture the full impact of genetic characteristics and lifelong factors on actual sleep patterns. Third, the extremely small sample size of individuals genetically classified as morning type (n=3) may have been insufficient to offset sampling bias.
However, this paradoxical finding may also reflect the mismatch between genetic chronotype predisposition and actual sleep-wake behavior in modern society, where social and work obligations may conflict with biological preferences. Our results align with recent GWAS identifying chronotype-associated variants and extend these findings by revealing the potential negative consequences of chronotype-environment misalignment. The observation that genetic morning types in our sample experienced sleep difficulties suggests that social jet lag, the discrepancy between biological and social time, may be particularly problematic for individuals whose genetic predisposition conflicts with their required sleep schedule [29].
Indeed, examination of sleep indicators across groups revealed minimal differences in sleep pattern-related measures (BT, LO, WT, and TOB) among the morning, intermediate, and evening chronotype groups, with all groups obtaining adequate sleep averaging 6–7 h. However, the disproportionately small morning-type group, with over half showing an average SOL and WASO exceeding 15 min, likely created blind spots in the group-based mean comparisons.
Genetic markers of calcium deficiency showed compensatory patterns, in which delayed sleep timing was accompa-nied by reduced daytime sleepiness. At the molecular level, calcium flux is essential for the rhythmic expression of Period 1 (Per1) in suprachiasmatic nucleus neurons, with calcium channel blockade completely abolishing the Per1 rhythm [6]. Reduced calcium availability may attenuate PER1 expression rhythms, contributing to evening chronotype tendencies. Unlike clinical calcium deficiency, which typically causes sleep disturbances, genetic calcium deficiency may promote a delayed but intrinsically aligned circadian phase. When individuals follow their genetically determined evening chronotype rather than fight against it, they achieve better circadian alignment, resulting in more consolidated sleep and reduced daytime sleepiness despite delayed timing.
This interpretation is supported by research showing that chronotype misalignment, rather than chronotype per se, is the primary driver of sleep problems [30]. Recent GWAS have identified variants in core clock genes, including PER1, as being associated with chronotype preferences. The calcium-PER1 pathway may represent a novel genetic mechanism underlying individual differences in chronotype [5].
CoQ10 deficiency markers demonstrated clear dose-response relationships, requiring extended sleep duration and delayed awakenings. CoQ10 deficiency may lead to mitochondrial dysfunction and cellular energy deficits requiring compensatory sleep extension to restore adequate energy levels. The NRXN1 gene’s involvement provides mechanistic insights, as this gene regulates synaptic plasticity and neurotransmitter release [31]—functions that are energy-dependent and particularly vulnerable to CoQ10 deficiency. These findings are consistent with those of CoQ10 supplementation studies showing sleep improvements, supporting the role of CoQ10 in sleep regulation via mitochondrial energy metabolism [32].
The nicotine metabolism findings validated the pharmacological predictions, with slower metabolizers showing worse sleep, even in nonsmoking populations. However, the observation that CYP2A6 CC carriers (increased metabolism) reported significantly higher morning refreshment than TT carriers (decreased metabolism) can be explained by the demographic characteristics of the study population. This study targeted healthy adults without smoking or alcohol problems with the screening criteria of Fagerström Test for Nicotine Dependence <6 for smoking and Alcohol Use Disorder Identification Test <8 for alcohol use disorder. Most participants reported minimal or no smoking. These characteristics suggest that the actual stimulant substance intake was equivalent across all genotype groups. When substance intake levels are equal but metabolic rates differ, groups with higher metabolic rates may experience less arousal before or during sleep, potentially leading to a higher reported morning refreshment.
Nicotine metabolism-related genetic markers have demonstrated robust associations with sleep quality. Our study demonstrated that participants with slower nicotine metabolism showed significantly worse sleep outcomes across multiple parameters: higher WASO in the “slow” group compared with “fast” and “medium” groups (MD=8.959 and 6.924 min, respectively; both p<0.05), lower SE (MD=3.318% and 2.256%, respectively; p<0.05), and higher PSQI scores indicating poorer overall SQ (MD=2.121; p=0.015).
These findings suggest that genetic variations affecting cytochrome P450 enzymes (particularly CYP2A6) may influence sleep through mechanisms beyond nicotine processing, potentially affecting the metabolism of endogenous sleepregulatory compounds or neurotransmitters [32-34]. Our study extends previous research on CYP2A6 polymorphisms and sleep by demonstrating that these genetic variations can predict sleep quality, even in nonsmoking populations, indicating broader metabolic implications for sleep regulation. The strong correlations observed in our sample (FRESH: r=-0.261, p<0.05; WASO: r=0.356, p<0.01) suggest that nicotine metabolism genetic markers could serve as valuable tools for identifying individuals at risk for sleep disorders, regardless of smoking status.
Gallbladder cancer markers showed significant associations with sleep initiation difficulties and reduced sleep efficiency [35]. Although the mechanistic link is unclear, this finding may reflect shared genetic pathways involving oxidative stress response or DNA repair mechanisms, as represented by the OGG1 polymorphism identified in our analyses. Our study is among the first to report associations between cancer susceptibility genes and sleep parameters in healthy populations. The strong correlations observed with SOL (r=0.341, p<0.01) and SE (r=-0.335, p<0.01) suggest that genetic variations affecting DNA repair mechanisms may have broader implications for sleep regulation than previously recognized.
To the best of our knowledge, this is the first study to demonstrate a direct association between genetic susceptibility to skin cancer and chronotype preference. Our findings revealed a remarkably consistent pattern across all three major skin cancer types, with genetic predisposition to basal cell carcinoma, malignant melanoma, and squamous cell carcinoma, all of which were associated with a morning chronotype preference and superior sleep quality.
The concordant findings across basal cell carcinoma (better PSQI scores, r=-0.249, p<0.05; morning chronotype, r=0.214, p<0.05), malignant melanoma (better PSQI scores, r=-0.242, p<0.05; higher MEQ scores, r=0.235, p<0.05), and squamous cell carcinoma (longer sleep duration with reduced daytime sleepiness, r=-0.215, p<0.05) suggest a shared underlying mechanism rather than gene-specific effects. This convergent pattern across genetically distinct cancer susceptibility pathways indicates that the relationship between predisposition to skin cancer and circadian timing may reflect fundamental biological trade-offs. Clear dose-response relationships further support the genetic causality of these associations, as demonstrated in malignant melanoma, where the “good” group showed superior sleep quality compared to the “warning” group (MD=0.799, p=0.002), and at the SNP level, where MTAP gene (rs7023329) AA carriers had significantly higher MEQ scores than GG carriers (MD=3.080, p=0.046), providing molecular-level evidence for the genetic basis of this chronotypecancer relationship.
This may reflect a fundamental trade-off in the genetic architecture between circadian alignment and UV-related cancer susceptibility, as described by the gene-environment interaction hypothesis. Morning-oriented individuals, with earlier wake times and potentially increased morning sunlight exposure, may face an elevated risk of skin cancer despite having superior sleep quality [36,37].
Although previous research has established that the morning chronotype is protective against breast cancer [38], our findings suggest that the relationship between chronotype and cancer risk may be tissue-specific and dependent on environmental exposure patterns. This discovery has important implications for personalized health recommendations, as individuals with genetic susceptibility to skin cancer may need to balance optimal circadian timing with UV protection strategies. Our findings provide novel evidence for this gene-environment interaction hypothesis, showing significant correlations between skin cancer markers and chronotype scores across multiple cancer types.
Clinical implications
These findings have important implications for personalized sleep medicine. Individuals with a genetic risk for calcium deficiency should receive chronotype-aligned interventions to optimize sleep within their preferred timing. Those with a CoQ10 deficiency genetic risk require treatment strategies focusing on sleep maintenance, whereas nicotine metabolism markers can identify individuals at risk for sleep disorders, regardless of smoking status. For individuals with gallbladder cancer-related markers who showed prolonged sleep-onset latency and reduced sleep efficiency, medications that facilitate sleep initiation may be more effective than those that maintain sleep. Individuals with genetic susceptibility to skin cancer require personalized prevention strategies to balance beneficial sleep patterns with enhanced UV protection. Therefore, treatment strategies that promote sleep onset should be prioritized for this population.
Complexity of gene-sleep interactions
Previous GWAS have revealed that many individual variants associated with sleep characteristics are located in regions overlapping genes related to circadian rhythms and neurotransmitters, as well as genes associated with cardiovascular and metabolic disease risk [39-41]. Some sleep-related genetic markers identified in previous studies were also found to consistently influence sleep phenotypes and sleep outcomes in this study (ERC2, AK5). Additionally, the results of this study show that mutations in genetic markers identified in previous studies did not determine sleep phenotypes in a simple manner. For instance, morning chronotype genetic markers were associated with longer SOL and lower SE (F=4.044 and 6.263, respectively; both p<0.05). In the present study, mutations in nicotine metabolism-related genes, which drive increased nicotine consumption, were associated with better sleep outcomes. These findings suggest that an individual’s genetic predisposition may ultimately affect sleep through interactions with the environment. Moreover, these paradoxes underscore the importance of considering genetic variants in their full phenotypic context rather than assuming uniform beneficial or detrimental effects across all health domains [42].
In addition, the discordance between cumulative genetic risk scores and individual SNP analyses demonstrates the complexity of gene-sleep interactions. Concordant results between Spearman’s correlations and group-level ANOVA were observed more frequently than in individual SNP-based analyses, suggesting that sleep phenotypes may be determined by the cumulative effects of multiple genetic factors rather than single variants. This finding emphasizes the importance of polygenic risk scoring approaches in personalized sleep medicine.
Our polygenic approach, which examined both health- and disease-related markers simultaneously, revealed intricate genesleep-disease networks that would not be apparent from single-gene studies. Although there are several limitations in interpreting the results, including sample size and participant demographics, these findings suggest that more diverse genetic markers should be considered in sleep-related polygenic risk score studies based on GWA data.
Methodological considerations and limitations
This study has several limitations. First, the relatively small sample size (n=85) and group imbalances limited the statistical power. Extreme group imbalances in some disease risk categories precluded ANOVA. The cross-sectional design precludes causal inferences regarding the relationships between genetic markers and sleep parameters [43]. Although our power analysis indicated an adequate sample size for detecting medium effect sizes (r=0.3) with 80% power, smaller effects may have been missed.
Second, the use of commercial genetic testing panels (Genestyle and Hellogene), while providing broad coverage across 230 health- and 317 disease-related markers, may not capture all relevant genetic variations affecting sleep. To address multiple testing concerns in analyzing approximately 600 genetic variables, we implemented stringent criteria requiring associations to be significant at p<0.01 and present in at least two sleep parameters.
Third, the generalizability of the findings may be limited due to the characteristics of our sample. The predominantly female sample (77.6%) may limit the generalizability, particularly given the known sex differences in sleep patterns and gene expression. The homogeneous population (Korean adults) limits the generalizability of our findings to other populations with different genetic backgrounds. In addition, focusing exclusively on healthy adults limits the generalizability to clinical populations.
Fourth, objective measurements for sleep-related variables were not provided. The use of self-reported sleep diaries and questionnaires could have biased the study’s findings.
Finally, although this study was conducted on healthy individuals, it is possible that participants with sleep or mental disorders were not completely excluded during the recruitment process. We collected participants’ medical histories, including sleep disorder symptoms based on the DSM-5, through telephone interviews to exclude individuals with physical, psychological, and sleep disorders. However, the participant recruitment process, which depended on subjective reports, had limitations in completely excluding potential disorders or symptoms. Therefore, future studies should be conducted on healthy individuals using more sophisticated designs.
Future research directions
Future studies should enhance the statistical power by using larger sample sizes and balanced group distributions. Longitudinal study designs are needed to clarify the causal relationships between genetic predispositions and sleep patterns and to develop comprehensive models that include environmental factor interactions. To implement personalized sleep medicine, predictive models that integrate polygenic risk scores with clinical phenotypes must be developed.
Investigating the mechanistic pathways underlying the observed associations is crucial for translating these findings into clinical applications. This includes examining the functional consequences of the identified genetic variants and their interactions with environmental factors.
Future studies should use both subjective and objective measurement tools to minimize potential biases when exploring the relationship between genetic markers and sleep patterns. Objective measurement of sleep indicators using wearable devices or polysomnography will help clarify the relationship between genetic predispositions and sleep patterns.
The development of polygenic risk scores incorporating multiple genetic variants associated with sleep parameters could enhance the clinical utility of genetic information in sleep medicine. Such approaches might provide more accurate risk predictions than single-marker analyses. Finally, randomized controlled clinical trials of genetically informed personalized treatment approaches are required to validate their clinical utility.
Conclusions
This study provides novel insights into the complex relationships between genetic markers and sleep parameters, revealing patterns that challenge conventional assumptions regarding health-related genetics.
Individuals with a genetic predisposition to calcium deficiency exhibit a remarkably adaptive pattern, showing delayed sleep timing but paradoxically reduced daytime sleepiness. This suggests that delayed but more consolidated sleep may serve as an effective compensatory mechanism for underlying calcium deficiency vulnerabilities. In contrast, those with a genetic predisposition to CoQ10 deficiency displayed compensatory sleep extension through longer sleep duration and delayed wake times, although this appeared to be an incomplete adaptation, as evidenced by poorer subjective sleep quality. The body appears to systematically require extra recovery time in response to the genetic vulnerability to CoQ10 deficiency.
The nicotine metabolism findings validated the pharmacological predictions, demonstrating that when smoking levels are similar, slower nicotine metabolism consistently predicts worse sleep quality. Remarkably, this pattern persisted even in predominantly nonsmoking populations, suggesting that cytochrome P450 enzymes may regulate endogenous sleep-related compounds beyond their traditional roles in drug metabolism. Surprisingly, genetic susceptibility to skin cancers consistently predicted superior sleep quality and morning chronotype preferences across all three cancer types, revealing a paradoxical relationship in which disease-risk genes confer sleep advantages.
These findings have profound clinical implications for personalized sleep medicine. Individuals with a genetic risk for calcium deficiency benefit from chronotype-aligned interventions that optimize sleep within their preferred later timing rather than forcing earlier schedules. Patients with a genetic risk for CoQ10 deficiency require treatment strategies that focus on sleep maintenance rather than sleep initiation. Nicotine metabolism markers can serve as valuable tools for identifying individuals at risk for sleep disorders, regardless of their smoking status. Most critically, individuals with a genetic susceptibility to skin cancer present a unique clinical scenario requiring personalized prevention strategies that balance their superior sleep quality and beneficial circadian alignment with enhanced UV protection counseling and more frequent dermatological screening.
The identification of genetic variants that simultaneously predict disease susceptibility and sleep advantages represents a new perspective in our understanding of health-related genetics, emphasizing that genetic architecture involves complex trade-offs between different physiological systems. Future research should focus on developing personalized treatment approaches that work with individual genetic predispositions, while establishing causality through larger longitudinal studies and creating clinically applicable polygenic risk scores.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Print 
