Genetic modifiers of rare variants in monogenic developmental disorder loci

Rare damaging variants in a large number of genes are known to cause monogenic developmental disorders (DDs) and have also been shown to cause milder subclinical phenotypes in population cohorts. Here, we show that carrying multiple (2-5) rare damaging variants across 599 dominant DD genes has an additive adverse effect on numerous cognitive and socioeconomic traits in UK Biobank, which can be partially counterbalanced by a higher educational attainment polygenic score (EA-PGS). Phenotypic deviators from expected EA-PGS could be partly explained by the enrichment or depletion of rare DD variants. Among carriers of rare DD variants, those with a DD-related clinical diagnosis had a substantially lower EA-PGS and more severe phenotype than those without a clinical diagnosis. Our results suggest that the overall burden of both rare and common variants can modify the expressivity of a phenotype, which may then influence whether an individual reaches the threshold for clinical disease.

Chromosome 20p11.2 deletions cause congenital hyperinsulinism via the loss of FOXA2 or its regulatory elements

Persistent congenital hyperinsulinism (HI) is a rare genetically heterogeneous condition characterised by dysregulated insulin secretion leading to life-threatening hypoglycaemia. For up to 50% of affected individuals screening of the known HI genes does not identify a disease-causing variant. Large deletions have previously been used to identify novel regulatory regions causing HI. Here, we used genome sequencing to search for novel large (>1 Mb) deletions in 180 probands with HI of unknown cause and replicated our findings in a large cohort of 883 genetically unsolved individuals with HI using off-target copy number variant calling from targeted gene panels. We identified overlapping heterozygous deletions in five individuals (range 3-8 Mb) spanning chromosome 20p11.2. The pancreatic beta-cell transcription factor gene, FOXA2, a known cause of HI was deleted in two of the five individuals. In the remaining three, we found a minimal deleted region of 2.4 Mb adjacent to FOXA2 that encompasses multiple non-coding regulatory elements that are in conformational contact with FOXA2. Our data suggests that the deletions in these three children may cause disease through the dysregulation of FOXA2 expression. These findings provide new insights into the regulation of FOXA2 in the beta-cell and confirm an aetiological role for chromosome 20p11.2 deletions in syndromic HI.

Applying a genetic risk score model to enhance prediction of future multiple sclerosis diagnosis at first presentation with optic neuritis

Optic neuritis (ON) is associated with numerous immune-mediated inflammatory diseases, but 50% patients are ultimately diagnosed with multiple sclerosis (MS). Differentiating MS-ON from non-MS-ON acutely is challenging but important; non-MS ON often requires urgent immunosuppression to preserve vision. Using data from the United Kingdom Biobank we showed that combining an MS-genetic risk score (GRS) with demographic risk factors (age, sex) significantly improved MS prediction in undifferentiated ON; one standard deviation of MS-GRS increased the Hazard of MS 1.3-fold (95% confidence interval 1.07-1.55, P < 0.01). Participants stratified into quartiles of predicted risk developed incident MS at rates varying from 4% (95%CI 0.5-7%, lowest risk quartile) to 41% (95%CI 33-49%, highest risk quartile). The model replicated across two cohorts (Geisinger, USA, and FinnGen, Finland). This study indicates that a combined model might enhance individual MS risk stratification, paving the way for precision-based ON treatment and earlier MS disease-modifying therapy.

Penetrance and expressivity of mitochondrial variants in a large clinically unselected population

Whole genome sequencing (WGS) from large clinically unselected cohorts provides a unique opportunity to assess the penetrance and expressivity of rare and/or known pathogenic mitochondrial variants in population. Using WGS from 179 862 clinically unselected individuals from the UK Biobank, we performed extensive single and rare variant aggregation association analyses of 15 881 mtDNA variants and 73 known pathogenic variants with 15 mitochondrial disease-relevant phenotypes. We identified 12 homoplasmic and one heteroplasmic variant (m.3243A>G) with genome-wide significant associations in our clinically unselected cohort. Heteroplasmic m.3243A>G (MAF = 0.0002, a known pathogenic variant) was associated with diabetes, deafness and heart failure and 12 homoplasmic variants increased aspartate aminotransferase levels including three low-frequency variants (MAF ~0.002 and beta~0.3 SD). Most pathogenic mitochondrial disease variants (n = 66/74) were rare in the population (<1:9000). Aggregated or single variant analysis of pathogenic variants showed low penetrance in unselected settings for the relevant phenotypes, except m.3243A>G. Multi-system disease risk and penetrance of diabetes, deafness and heart failure greatly increased with m.3243A>G level ≥ 10%. The odds ratio of these traits increased from 5.61, 12.3 and 10.1 to 25.1, 55.0 and 39.5, respectively. Diabetes risk with m.3243A>G was further influenced by type 2 diabetes genetic risk. Our study of mitochondrial variation in a large-unselected population identified novel associations and demonstrated that pathogenic mitochondrial variants have lower penetrance in clinically unselected settings. m.3243A>G was an exception at higher heteroplasmy showing a significant impact on health making it a good candidate for incidental reporting.

Genome-wide association analysis of composite sleep health scores in 413,904 individuals

Recent genome-wide association studies (GWASs) of several individual sleep traits have identified hundreds of genetic loci, suggesting diverse mechanisms. Moreover, sleep traits are moderately correlated, and together may provide a more complete picture of sleep health, while also illuminating distinct domains. Here we construct novel sleep health scores (SHSs) incorporating five core self-report measures: sleep duration, insomnia symptoms, chronotype, snoring, and daytime sleepiness, using additive (SHS-ADD) and five principal components-based (SHS-PCs) approaches. GWASs of these six SHSs identify 28 significant novel loci adjusting for multiple testing on six traits (p<8.3e-9), along with 341 previously reported loci (p<5e-08). The heritability of the first three SHS-PCs equals or exceeds that of SHS-ADD (SNP-h2=0.094), while revealing sleep-domain-specific genetic discoveries. Significant loci enrich in multiple brain tissues and in metabolic and neuronal pathways. Post GWAS analyses uncover novel genetic mechanisms underlying sleep health and reveal connections to behavioral, psychological, and cardiometabolic traits.

Hyperglycaemia is a causal risk factor for upper limb pathologies

BACKGROUND

Diabetes (regardless of type) and obesity are associated with a range of musculoskeletal disorders. The causal mechanisms driving these associations are unknown for many upper limb pathologies. We used genetic techniques to test the causal link between glycemia, obesity and musculoskeletal conditions.

METHODS

In the UK Biobank's unrelated European cohort (N = 379 708) we performed mendelian randomisation (MR) analyses to test for a causal effect of long-term high glycaemia and adiposity on four musculoskeletal pathologies: frozen shoulder, Dupuytren's disease, carpal tunnel syndrome and trigger finger. We also performed single-gene MR using rare variants in the GCK gene.

RESULTS

Using MR, we found evidence that long-term high glycaemia has a causal role in the aetiology of upper limb conditions. A 10-mmol/mol increase in genetically predicted haemoglobin A1C (HbA1c) was associated with frozen shoulder: odds ratio (OR) = 1.50 [95% confidence interval (CI), 1.20-1.88], Dupuytren's disease: OR = 1.17 (95% CI, 1.01-1.35), trigger finger: OR = 1.30 (95% CI, 1.09-1.55) and carpal tunnel syndrome: OR = 1.20 (95% CI, 1.09-1.33). Carriers of GCK mutations have increased odds of frozen shoulder: OR = 7.16 (95% CI, 2.93-17.51) and carpal tunnel syndrome: OR = 2.86 (95% CI, 1.50-5.44) but not Dupuytren's disease or trigger finger. We found evidence that an increase in genetically predicted body mass index (BMI) of 5 kg/m2 was associated with carpal tunnel syndrome: OR = 1.13 (95% CI, 1.10-1.16) and associated negatively with Dupuytren's disease: OR = 0.94 (95% CI, 0.90-0.98), but no evidence of association with frozen shoulder or trigger finger. Trigger finger (OR 1.96 (95% CI, 1.42-2.69) P = 3.6e-05) and carpal tunnel syndrome [OR 1.63 (95% CI, 1.36-1.95) P = 8.5e-08] are associated with genetically predicted unfavourable adiposity increase of one standard deviation of body fat.

CONCLUSIONS

Our study consistently demonstrates a causal role of long-term high glycaemia in the aetiology of upper limb musculoskeletal conditions. Clinicians treating diabetes patients should be aware of these complications in clinic, specifically those managing the care of GCK mutation carriers. Upper limb musculoskeletal conditions should be considered diabetes complications.

Exploring the application of deep learning methods for polygenic risk score estimation

Background

Polygenic risk scores (PRS) summarise genetic information into a single number with multiple clinical and research uses. Machine learning (ML) has revolutionised a diverse set of fields, however, the impact of ML on genomics in general, and PRSs in particular, has been less significant. We explore how ML can improve the generation of PRSs.

Methods

We train ML models on known PRSs using UK Biobank data. We explore whether the models can recreate human programmed PRSs, including using a single model to generate multiple PRSs, and the difficulty in using ML for PRS generation. We also investigate how ML can compensate for missing data and the constraints on performance.

Results

We demonstrate almost perfect generation of PRSs, including when using one model to predict multiple scores, and with little loss of performance with reduced quantity of training data. For an example set of missing SNPs the MLP produces predictions that enable separation of cases from population samples with an area under the receiver operating characteristic curve of 0.847 (95% CI: 0.828-0.864) compared to 0.798 (95% CI: 0.779-0.818) for the PRS. We provide evidence that input information is the limiting factor of further improvement.

Conclusions

ML can accurately generate PRSs, including with one model for multiple PRSs. The models are transferable and have high longevity. With certain missing SNPs the ML models can statistically significantly improve on normal PRS generation. Models trained are probably at the edge of performance and further improvements likely require use of additional input data.

Correlations in sleeping patterns and circadian preference between spouses

Spouses may affect each other's sleeping behaviour. In 47,420 spouse-pairs from the UK Biobank, we found a weak positive phenotypic correlation between spouses for self-reported sleep duration (r = 0.11; 95% CI = 0.10, 0.12) and a weak inverse correlation for chronotype (diurnal preference) (r = -0.11; -0.12, -0.10), which replicated in up to 127,035 23andMe spouse-pairs. Using accelerometer data on 3454 UK Biobank spouse-pairs, the correlation for derived sleep duration was similar to self-report (r = 0.12; 0.09, 0.15). Timing of diurnal activity was positively correlated (r = 0.24; 0.21, 0.27) in contrast to the inverse correlation for chronotype. In Mendelian randomization analysis, positive effects of sleep duration (mean difference=0.13; 0.04, 0.23 SD per SD) and diurnal activity (0.49; 0.03, 0.94) were observed, as were inverse effects of chronotype (-0.15; -0.26, -0.04) and snoring (-0.15; -0.27, -0.04). Findings support the notion that an individual's sleep may impact that of their partner, promoting opportunities for sleep interventions at the family-level.

The Role of ONECUT1 Variants in Monogenic and Type 2 Diabetes Mellitus

ONECUT1 (also known as HNF6) is a transcription factor involved in pancreatic development and β-cell function. Recently, biallelic variants in ONECUT1 were reported as a cause of neonatal diabetes mellitus (NDM) in two subjects, and missense monoallelic variants were associated with type 2 diabetes and possibly maturity-onset diabetes of the young (MODY). Here we examine the role of ONECUT1 variants in NDM, MODY, and type 2 diabetes in large international cohorts of subjects with monogenic diabetes and >400,000 subjects from UK Biobank. We identified a biallelic frameshift ONECUT1 variant as the cause of NDM in one individual. However, we found no enrichment of missense or null ONECUT1 variants among 484 individuals clinically suspected of MODY, in whom all known genes had been excluded. Finally, using a rare variant burden test in the UK Biobank European cohort, we identified a significant association between heterozygous ONECUT1 null variants and type 2 diabetes (P = 0.006) but did not find an association between missense variants and type 2 diabetes. Our results confirm biallelic ONECUT1 variants as a cause of NDM and highlight monoallelic null variants as a risk factor for type 2 diabetes. These findings confirm the critical role of ONECUT1 in human β-cell function.

ARTICLE HIGHLIGHTS

Insights into the genetics of menopausal vasomotor symptoms: genome-wide analyses of routinely-collected primary care health records

BACKGROUND

Vasomotor symptoms (VMS) can often significantly impact women's quality of life at menopause. In vivo studies have shown that increased neurokinin B (NKB) / neurokinin 3 receptor (NK3R) signalling contributes to VMS, with previous genetic studies implicating the TACR3 gene locus that encodes NK3R. Large-scale genomic analyses offer the possibility of biological insights but few such studies have collected data on VMS, while proxy phenotypes such as hormone replacement therapy (HRT) use are likely to be affected by changes in clinical practice. We investigated the genetic basis of VMS by analysing routinely-collected health records.

METHODS

We performed a GWAS of VMS derived from linked primary-care records and cross-sectional self-reported HRT use in up to 153,152 women from UK Biobank, a population-based cohort. In a subset of this cohort (n = 39,356), we analysed exome-sequencing data to test the association with VMS of rare deleterious genetic variants. Finally, we used Mendelian randomisation analysis to investigate the reasons for HRT use over time.

RESULTS

Our GWAS of health-records derived VMS identified a genetic signal near TACR3 associated with a lower risk of VMS (OR=0.76 (95% CI 0.72,0.80) per A allele, P=3.7x10-27), which was consistent with previous studies, validating this approach. Conditional analyses demonstrated independence of genetic signals for puberty timing and VMS at the TACR3 locus, including a rare variant predicted to reduce functional NK3R levels that was associated with later menarche (P = 5 × 10-9) but showed no association with VMS (P = 0.6). Younger menopause age was causally-associated with greater HRT use before 2002 but not after.

CONCLUSIONS

We provide support for TACR3 in the genetic basis of VMS but unexpectedly find that rare genomic variants predicted to lower NK3R levels did not modify VMS, despite the proven efficacy of NK3R antagonists. Using genomics we demonstrate changes in genetic associations with HRT use over time, arising from a change in clinical practice since the early 2000s, which is likely to reflect a switch from preventing post-menopausal complications in women with earlier menopause to primarily treating VMS. Our study demonstrates that integrating routinely-collected primary care health records and genomic data offers great potential for exploring the genetic basis of symptoms.

Influence of family history on penetrance of hereditary cancers in a population setting

Background

We sought to investigate how penetrance of familial cancer syndromes varies with family history using a population-based cohort.

Methods

We analysed 454,712 UK Biobank participants with exome sequence and clinical data (data collected between March 2006 and June 2021). We identified participants with a self-reported family history of breast or colorectal cancer and a pathogenic/likely pathogenic variant in the major genes responsible for hereditary breast cancer or Lynch syndrome. We calculated survival to cancer diagnosis (controlled for sex, death, recruitment centre, screening and prophylactic surgery).

Findings

Women with a pathogenic BRCA1 or BRCA2 variant had an increased risk of breast cancer that was higher in those with a first-degree family history (relative hazard 10.3 and 7.8, respectively) than those without (7.2 and 4.7). Penetrance to age 60 was also higher in those with a family history (44.7%, CI 32.2-59.3 and 24.1%, CI 17.5-32.6) versus those without (22.8%, CI 15.9-32.0 and 17.9%, CI 13.8-23.0). A similar pattern was seen in Lynch syndrome: individuals with a pathogenic MLH1, MSH2 or MSH6 variant had an increased risk of colorectal cancer that was significantly higher in those with a family history (relative hazard 35.6, 48.0 and 9.9) than those without (13.0, 15.4 and 7.2). Penetrance to age 60 was also higher for carriers of a pathogenic MLH1 or MSH2 variant in those with a family history (30.9%, CI 18.1-49.3 and 38.3%, CI 21.5-61.8) versus those without (20.5% CI 9.6-40.5 and 8.3% CI 2.1-30.4), but not for MSH6 (6.5% CI 2.7-15.1 with family history versus 8.3%, CI 5.1-13.2). Relative risk increases were also observed both within and across conditions.

Interpretation

Individuals with pathogenic cancer syndrome variants may be at a less elevated risk of cancer in the absence of a first-degree family history, so in the context of results return, family history should be considered when counselling patients on the risks and benefits of potential follow-up care.

Funding

The current work is supported by the MRC (grant no MR/T00200X/1). The MRC had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Type 1 Diabetes Genetic Risk Score Differentiates Subgroups of Ketosis-Prone Diabetes

OBJECTIVE

To determine whether genetic risk for type 1 diabetes (T1D) differentiates the four Aβ subgroups of ketosis-prone diabetes (KPD), where A+ and A- define the presence or absence of islet autoantibodies and β+ and β- define the presence or absence of β-cell function.

RESEARCH DESIGN AND METHODS

We compared T1D genetic risk scores (GRS) of patients with KPD across subgroups, race/ethnicity, β-cell function, and glycemia.

RESULTS

Among 426 patients with KPD (54% Hispanic, 31% African American, 11% White), rank order of GRS was A+β- > A+β+ = A-β- > A-β+. GRS of A+β- KPD was lower than that of a T1D cohort, and GRS of A-β+ KPD was higher than that of a type 2 diabetes cohort. GRS was lowest among African American patients, with a similar distribution across KPD subgroups.

CONCLUSIONS

T1D genetic risk delineates etiologic differences among KPD subgroups. Patients with A+β- KPD have the highest and those with A-β+ KPD the lowest GRS.

Identification and analysis of individuals who deviate from their genetically-predicted phenotype

Findings from genome-wide association studies have facilitated the generation of genetic predictors for many common human phenotypes. Stratifying individuals misaligned to a genetic predictor based on common variants may be important for follow-up studies that aim to identify alternative causal factors. Using genome-wide imputed genetic data, we aimed to classify 158,951 unrelated individuals from the UK Biobank as either concordant or deviating from two well-measured phenotypes. We first applied our methods to standing height: our primary analysis classified 244 individuals (0.15%) as misaligned to their genetically predicted height. We show that these individuals are enriched for self-reporting being shorter or taller than average at age 10, diagnosed congenital malformations, and rare loss-of-function variants in genes previously catalogued as causal for growth disorders. Secondly, we apply our methods to LDL cholesterol (LDL-C). We classified 156 (0.12%) individuals as misaligned to their genetically predicted LDL-C and show that these individuals were enriched for both clinically actionable cardiovascular risk factors and rare genetic variants in genes previously shown to be involved in metabolic processes. Individuals whose LDL-C was higher than expected based on the genetic predictor were also at higher risk of developing coronary artery disease and type-two diabetes, even after adjustment for measured LDL-C, BMI and age, suggesting upward deviation from genetically predicted LDL-C is indicative of generally poor health. Our results remained broadly consistent when performing sensitivity analysis based on a variety of parametric and non-parametric methods to define individuals deviating from polygenic expectation. Our analyses demonstrate the potential importance of quantitatively identifying individuals for further follow-up based on deviation from genetic predictions.

Utility of genetic risk scores in type 1 diabetes

Iterative advances in understanding of the genetics of type 1 diabetes have identified >70 genetic regions associated with risk of the disease, including strong associations across the HLA class II region that account for >50% of heritability. The increased availability of genetic data combined with the decreased costs of generating these data, have facilitated the development of polygenic scores that aggregate risk variants from associated loci into a single number: either a genetic risk score (GRS) or a polygenic risk score (PRS). PRSs incorporate the risk of many possibly correlated variants from across the genome, even if they do not reach genome-wide significance, whereas GRSs estimate the cumulative contribution of a smaller subset of genetic variants that reach genome-wide significance. Type 1 diabetes GRSs have utility in diabetes classification, aiding discrimination between type 1 diabetes, type 2 diabetes and MODY. Type 1 diabetes GRSs are also being used in newborn screening studies to identify infants at risk of future presentation of the disease. Most early studies of type 1 diabetes genetics have been conducted in European ancestry populations, but, to develop accurate GRSs across diverse ancestries, large case-control cohorts from non-European populations are still needed. The current barriers to GRS implementation within healthcare are mainly related to a lack of guidance and knowledge on integration with other biomarkers and clinical variables. Once these limitations are addressed, there is huge potential for 'test and treat' approaches to be used to tailor care for individuals with type 1 diabetes.

Associations between sleep health and grey matter volume in the UK Biobank cohort (n = 33 356)

As suggested by previous research, sleep health is assumed to be a key determinant of future morbidity and mortality. In line with this, recent studies have found that poor sleep is associated with impaired cognitive function. However, to date, little is known about brain structural abnormalities underlying this association. Although recent findings link sleep health deficits to specific alterations in grey matter volume, evidence remains inconsistent and reliant on small sample sizes. Addressing this problem, the current preregistered study investigated associations between sleep health and grey matter volume (139 imaging-derived phenotypes) in the UK Biobank cohort (33 356 participants). Drawing on a large sample size and consistent data acquisition, sleep duration, insomnia symptoms, daytime sleepiness, chronotype, sleep medication and sleep apnoea were examined. Our main analyses revealed that long sleep duration was systematically associated with larger grey matter volume of basal ganglia substructures. Insomnia symptoms, sleep medication and sleep apnoea were not associated with any of the 139 imaging-derived phenotypes. Short sleep duration, daytime sleepiness as well as late and early chronotype were associated with solitary imaging-derived phenotypes (no recognizable pattern, small effect sizes). To our knowledge, this is the largest study to test associations between sleep health and grey matter volume. Clinical implications of the association between long sleep duration and larger grey matter volume of basal ganglia are discussed. Insomnia symptoms as operationalized in the UK Biobank do not translate into grey matter volume findings.

Penetrance of pathogenic genetic variants associated with premature ovarian insufficiency

Premature ovarian insufficiency (POI) affects 1% of women and is a leading cause of infertility. It is often considered to be a monogenic disorder, with pathogenic variants in ~100 genes described in the literature. We sought to systematically evaluate the penetrance of variants in these genes using exome sequence data in 104,733 women from the UK Biobank, 2,231 (1.14%) of whom reported at natural menopause under the age of 40 years. We found limited evidence to support any previously reported autosomal dominant effect. For nearly all heterozygous effects on previously reported POI genes, we ruled out even modest penetrance, with 99.9% (13,699 out of 13,708) of all protein-truncating variants found in reproductively healthy women. We found evidence of haploinsufficiency effects in several genes, including TWNK (1.54 years earlier menopause, P = 1.59 × 10-6) and SOHLH2 (3.48 years earlier menopause, P = 1.03 × 10-4). Collectively, our results suggest that, for the vast majority of women, POI is not caused by autosomal dominant variants either in genes previously reported or currently evaluated in clinical diagnostic panels. Our findings, plus previous studies, suggest that most POI cases are likely oligogenic or polygenic in nature, which has important implications for future clinical genetic studies, and genetic counseling for families affected by POI.

Prevalence of Fabry disease-causing variants in the UK Biobank

BACKGROUND

Fabry disease is an X-linked lysosomal storage disorder resulting from deficiency of the alpha-galactosidase A enzyme leading to accumulation of globotriaosylceramide in multiple organ sites with prominent cardiovascular and renal involvement. Global prevalence estimates of Fabry disease based on clinical ascertainment range from 1 in 40 000 to 1 in 170 000. We aimed to determine the prevalence of Fabry disease-causing variants in UK Biobank.

METHODS

We sought GLA gene variants in exome sequencing data from 200 643 individuals from UK Biobank. We used ACMG/AMP guidelines (American College of Medical Genetics/Association for Molecular Pathology) to classify pathogenicity and compared baseline biomarker data, hospital ICD-10 (International Classification of Diseases version-10) codes, general practitioner records and self-reported health data with those without pathogenic variants.

RESULTS

We identified 81 GLA coding variants. We identified eight likely pathogenic variants on the basis of being rare (<1/10 000 individuals) and either previously reported to cause Fabry disease, or being protein-truncating variants. Thirty-six individuals carried one of these variants. In the UK Biobank, the prevalence of likely pathogenic Fabry disease-causing variants is 1/5732 for late-onset disease-causing variants and 1/200 643 for variants causing classic Fabry disease.

CONCLUSION

Fabry disease-causing GLA variants are more prevalent in an unselected population sample than the reported prevalence of Fabry disease. These are overwhelmingly variants associated with later onset. It is possible the prevalence of later-onset Fabry disease exceeds current estimates.

A founder UMOD variant is a common cause of hereditary nephropathy in the British population

BACKGROUND

Monogenic disorders are estimated to account for 10%-12% of patients with kidney failure. We report the unexpected finding of an unusual uromodulin (UMOD) variant in multiple pedigrees within the British population and demonstrate a shared haplotype indicative of an ancestral variant.

METHODS

Probands from 12 apparently unrelated pedigrees with a family history of kidney failure within a geographically contiguous UK region were shown to be heterozygous for a pathogenic variant of UMOD c.278_289delTCTGCCCCGAAG insCCGCCTCCT.

RESULTS

A total of 88 clinically affected individuals were identified, all born in the UK and of white British ethnicity. 20 other individuals with the variant were identified in the UK 100,000 Genomes (100K) Project and 9 from UK Biobank (UKBB). A common extended haplotype was present in 5 of the UKBB individuals who underwent genome sequencing which was only present in <1 in 5000 of UKBB controls. Significantly, rare variants (<1 in 250 general population) identified within 1 Mb of the UMOD variant by genome sequencing were detected in all of the 100K individuals, indicative of an extended shared haplotype.

CONCLUSION

Our data confirm a likely founder UMOD variant with a wide geographical distribution within the UK. It should be suspected in cases of unexplained familial nephropathy presenting in patients of white British ancestry.

Identification and analysis of individuals who deviate from their genetically-predicted phenotype

Findings from genome-wide association studies have facilitated the generation of genetic predictors for many common human phenotypes. Stratifying individuals misaligned to a genetic predictor based on common variants may be important for follow-up studies that aim to identify alternative causal factors. Using genome-wide imputed genetic data, we aimed to classify 158,951 unrelated individuals from the UK Biobank as either concordant or deviating from two well-measured phenotypes. We first applied our methods to standing height: our primary analysis classified 244 individuals (0.15%) as misaligned to their genetically predicted height. We show that these individuals are enriched for self-reporting being shorter or taller than average at age 10, diagnosed congenital malformations, and rare loss-of-function variants in genes previously catalogued as causal for growth disorders. Secondly, we apply our methods to LDL cholesterol. We classified 156 (0.12%) individuals as misaligned to their genetically predicted LDL cholesterol and show that these individuals were enriched for both clinically actionable cardiovascular risk factors and rare genetic variants in genes previously shown to be involved in metabolic processes. Individuals whose LDL-C was higher than expected based on the genetic predictor were also at higher risk of developing coronary artery disease and type-two diabetes, even after adjustment for measured LDL-C, BMI and age, suggesting upward deviation from genetically predicted LDL-C is indicative of generally poor health. Our results remained broadly consistent when performing sensitivity analysis based on a variety of parametric and non-parametric methods to define individuals deviating from polygenic expectation. Our analyses demonstrate the potential importance of quantitatively identifying individuals for further follow-up based on deviation from genetic predictions.

Author Summary

Human genetics is becoming increasingly useful to help predict human traits across a population owing to findings from large-scale genetic association studies and advances in the power of genetic predictors. This provides an opportunity to potentially identify individuals that deviate from genetic predictions for a common phenotype under investigation. For example, an individual may be genetically predicted to be tall, but be shorter than expected. It is potentially important to identify individuals who deviate from genetic predictions as this can facilitate further follow-up to assess likely causes. Using 158,951 unrelated individuals from the UK Biobank, with height and LDL cholesterol, as exemplar traits, we demonstrate that approximately 0.15% & 0.12% of individuals deviate from their genetically predicted phenotypes respectively. We observed these individuals to be enriched for a range of rare clinical diagnoses, as well as rare genetic factors that may be causal. Our analyses also demonstrate several methods for detecting individuals who deviate from genetic predictions that can be applied to a range of continuous human phenotypes.

The relationship between islet autoantibody status and the genetic risk of type 1 diabetes in adult-onset type 1 diabetes

AIMS/HYPOTHESIS

The reason for the observed lower rate of islet autoantibody positivity in clinician-diagnosed adult-onset vs childhood-onset type 1 diabetes is not known. We aimed to explore this by assessing the genetic risk of type 1 diabetes in autoantibody-negative and -positive children and adults.

METHODS

We analysed GAD autoantibodies, insulinoma-2 antigen autoantibodies and zinc transporter-8 autoantibodies (ZnT8A) and measured type 1 diabetes genetic risk by genotyping 30 type 1 diabetes-associated variants at diagnosis in 1814 individuals with clinician-diagnosed type 1 diabetes (1112 adult-onset, 702 childhood-onset). We compared the overall type 1 diabetes genetic risk score (T1DGRS) and non-HLA and HLA (DR3-DQ2, DR4-DQ8 and DR15-DQ6) components with autoantibody status in those with adult-onset and childhood-onset diabetes. We also measured the T1DGRS in 1924 individuals with type 2 diabetes from the Wellcome Trust Case Control Consortium to represent non-autoimmune diabetes control participants.

RESULTS

The T1DGRS was similar in autoantibody-negative and autoantibody-positive clinician-diagnosed childhood-onset type 1 diabetes (mean [SD] 0.274 [0.034] vs 0.277 [0.026], p=0.4). In contrast, the T1DGRS in autoantibody-negative adult-onset type 1 diabetes was lower than that in autoantibody-positive adult-onset type 1 diabetes (mean [SD] 0.243 [0.036] vs 0.271 [0.026], p<0.0001) but higher than that in type 2 diabetes (mean [SD] 0.229 [0.034], p<0.0001). Autoantibody-negative adults were more likely to have the more protective HLA DR15-DQ6 genotype (15% vs 3%, p<0.0001), were less likely to have the high-risk HLA DR3-DQ2/DR4-DQ8 genotype (6% vs 19%, p<0.0001) and had a lower non-HLA T1DGRS (p<0.0001) than autoantibody-positive adults. In contrast to children, autoantibody-negative adults were more likely to be male (75% vs 59%), had a higher BMI (27 vs 24 kg/m²) and were less likely to have other autoimmune conditions (2% vs 10%) than autoantibody-positive adults (all p<0.0001). In both adults and children, type 1 diabetes genetic risk was unaffected by the number of autoantibodies (p>0.3). These findings, along with the identification of seven misclassified adults with monogenic diabetes among autoantibody-negative adults and the results of a sensitivity analysis with and without measurement of ZnT8A, suggest that the intermediate type 1 diabetes genetic risk in autoantibody-negative adults is more likely to be explained by the inclusion of misclassified non-autoimmune diabetes (estimated to represent 67% of all antibody-negative adults, 95% CI 61%, 73%) than by the presence of unmeasured autoantibodies or by a discrete form of diabetes. When these estimated individuals with non-autoimmune diabetes were adjusted for, the prevalence of autoantibody positivity in adult-onset type 1 diabetes was similar to that in children (93% vs 91%, p=0.4).

CONCLUSIONS/INTERPRETATION

The inclusion of non-autoimmune diabetes is the most likely explanation for the observed lower rate of autoantibody positivity in clinician-diagnosed adult-onset type 1 diabetes. Our data support the utility of islet autoantibody measurement in clinician-suspected adult-onset type 1 diabetes in routine clinical practice.

Identifying type 1 and 2 diabetes in research datasets where classification biomarkers are unavailable: assessing the accuracy of published approaches

OBJECTIVES

We aimed to compare the performance of approaches for classifying insulin-treated diabetes within research datasets without measured classification biomarkers, evaluated against two independent biological definitions of diabetes type.

STUDY DESIGN AND SETTING

We compared accuracy of ten reported approaches for classifying insulin-treated diabetes into type 1 (T1D) and type 2 (T2D) diabetes in two cohorts: UK Biobank (UKBB) n = 26,399 and Diabetes Alliance for Research in England (DARE) n = 1,296. The overall performance for classifying T1D and T2D was assessed using: a T1D genetic risk score and genetic stratification method (UKBB); C-peptide measured at >3 years diabetes duration (DARE).

RESULTS

Approaches' accuracy ranged from 71% to 88% (UKBB) and 68% to 88% (DARE). When classifying all participants, combining early insulin requirement with a T1D probability model (incorporating diagnosis age and body image issue [BMI]), and interview-reported diabetes type (UKBB available in only 15%) consistently achieved high accuracy (UKBB 87% and 87% and DARE 85% and 88%, respectively). For identifying T1D with minimal misclassification, models with high thresholds or young diagnosis age (<20 years) had highest performance. Findings were incorporated into an online tool identifying optimum approaches based on variable availability.

CONCLUSION

Models combining continuous features with early insulin requirement are the most accurate methods for classifying insulin-treated diabetes in research datasets without measured classification biomarkers.

Reduced penetrance of MODY-associated HNF1A/HNF4A variants but not GCK variants in clinically unselected cohorts

The true prevalence and penetrance of monogenic disease variants are often not known because of clinical-referral ascertainment bias. We comprehensively assess the penetrance and prevalence of pathogenic variants in HNF1A, HNF4A, and GCK that account for >80% of monogenic diabetes. We analyzed clinical and genetic data from 1,742 clinically referred probands, 2,194 family members, clinically unselected individuals from a US health system-based cohort (n = 132,194), and a UK population-based cohort (n = 198,748). We show that one in 1,500 individuals harbor a pathogenic variant in one of these genes. The penetrance of diabetes for HNF1A and HNF4A pathogenic variants was substantially lower in the clinically unselected individuals compared to clinically referred probands and was dependent on the setting (32% in the population, 49% in the health system cohort, 86% in a family member, and 98% in probands for HNF1A). The relative risk of diabetes was similar across the clinically unselected cohorts highlighting the role of environment/other genetic factors. Surprisingly, the penetrance of pathogenic GCK variants was similar across all cohorts (89%-97%). We highlight that pathogenic variants in HNF1A, HNF4A, and GCK are not ultra-rare in the population. For HNF1A and HNF4A, we need to tailor genetic interpretation and counseling based on the setting in which a pathogenic monogenic variant was identified. GCK is an exception with near-complete penetrance in all settings. This along with the clinical implication of diagnosis makes it an excellent candidate for the American College of Medical Genetics secondary gene list.

Recurrent 17q12 microduplications contribute to renal disease but not diabetes

BACKGROUND

17q12 microdeletion and microduplication syndromes present as overlapping, multisystem disorders. We assessed the disease phenotypes of individuals with 17q12 CNV in a population-based cohort.

METHODS

We investigated 17q12 CNV using microarray data from 450 993 individuals in the UK Biobank and calculated disease status associations for diabetes, liver and renal function, neurological and psychiatric traits.

RESULTS

We identified 11 17q12 microdeletions and 106 microduplications. Microdeletions were strongly associated with diabetes (p=2×10^-7) but microduplications were not. Estimated glomerular filtration rate (eGFR mL/min/1.73 m²) was consistently lower in individuals with microdeletions (p=3×10^-12) and microduplications (p=6×10^-25). Similarly, eGFR <60, including end-stage renal disease, was associated with microdeletions (p=2×10^-9, p<0.003) and microduplications (p=1×10^-9, p=0.009), respectively, highlighting sometimes substantially reduced renal function in each. Microduplications were associated with decreased fluid intelligence (p=3×10^-4). SNP association analysis in the 17q12 region implicated changes to HNF1B as causing decreased eGFR (NC_000017.11:g.37741642T>G, rs12601991, p=4×10^-21) and diabetes (NC_000017.11:g.37741165C>T, rs7501939, p=6×10^-17). A second locus within the region was also associated with fluid intelligence (NC_000017.11:g.36593168T>C, rs1005552, p=6×10^-9) and decreased eGFR (NC_000017.11:g.36558947T>C, rs12150665, p=4×10^-15).

CONCLUSION

We demonstrate 17q12 microdeletions but not microduplications are associated with diabetes in a population-based cohort, likely caused by HNF1B haploinsufficiency. We show that both 17q12 microdeletions and microduplications are associated with renal disease, and multiple genes within the region likely contribute to renal and neurocognitive phenotypes.

The impact of Mendelian sleep and circadian genetic variants in a population setting

Rare variants in ten genes have been reported to cause Mendelian sleep conditions characterised by extreme sleep duration or timing. These include familial natural short sleep (ADRB1, DEC2/BHLHE41, GRM1 and NPSR1), advanced sleep phase (PER2, PER3, CRY2, CSNK1D and TIMELESS) and delayed sleep phase (CRY1). The association of variants in these genes with extreme sleep conditions were usually based on clinically ascertained families, and their effects when identified in the population are unknown. We aimed to determine the effects of these variants on sleep traits in large population-based cohorts. We performed genetic association analysis of variants previously reported to be causal for Mendelian sleep and circadian conditions. Analyses were performed using 191,929 individuals with data on sleep and whole-exome or genome-sequence data from 4 population-based studies: UK Biobank, FINRISK, Health-2000-2001, and the Multi-Ethnic Study of Atherosclerosis (MESA). We identified sleep disorders from self-report, hospital and primary care data. We estimated sleep duration and timing measures from self-report and accelerometery data. We identified carriers for 10 out of 12 previously reported pathogenic variants for 8 of the 10 genes. They ranged in frequency from 1 individual with the variant in CSNK1D to 1,574 individuals with a reported variant in the PER3 gene in the UK Biobank. No carriers for variants reported in NPSR1 or PER2 were identified. We found no association between variants analyzed and extreme sleep or circadian phenotypes. Using sleep timing as a proxy measure for sleep phase, only PER3 and CRY1 variants demonstrated association with earlier and later sleep timing, respectively; however, the magnitude of effect was smaller than previously reported (sleep midpoint ~7 mins earlier and ~5 mins later, respectively). We also performed burden tests of protein truncating (PTVs) or rare missense variants for the 10 genes. Only PTVs in PER2 and PER3 were associated with a relevant trait (for example, 64 individuals with a PTV in PER2 had an odds ratio of 4.4 for being "definitely a morning person", P = 4x10-8; and had a 57-minute earlier midpoint sleep, P = 5x10-7). Our results indicate that previously reported variants for Mendelian sleep and circadian conditions are often not highly penetrant when ascertained incidentally from the general population.

Detection and characterization of male sex chromosome abnormalities in the UK Biobank study

PURPOSE

The study aimed to systematically ascertain male sex chromosome abnormalities, 47,XXY (Klinefelter syndrome [KS]) and 47,XYY, and characterize their risks of adverse health outcomes.

METHODS

We analyzed genotyping array or exome sequence data in 207,067 men of European ancestry aged 40 to 70 years from the UK Biobank and related these to extensive routine health record data.

RESULTS

Only 49 of 213 (23%) of men whom we identified with KS and only 1 of 143 (0.7%) with 47,XYY had a diagnosis of abnormal karyotype on their medical records or self-report. We observed expected associations for KS with reproductive dysfunction (late puberty: risk ratio [RR] = 2.7; childlessness: RR = 4.2; testosterone concentration: RR = -3.8 nmol/L, all P < 2 × 10-8), whereas XYY men appeared to have normal reproductive function. Despite this difference, we identified several higher disease risks shared across both KS and 47,XYY, including type 2 diabetes (RR = 3.0 and 2.6, respectively), venous thrombosis (RR = 6.4 and 7.4, respectively), pulmonary embolism (RR = 3.3 and 3.7, respectively), and chronic obstructive pulmonary disease (RR = 4.4 and 4.6, respectively) (all P < 7 × 10-6).

CONCLUSION

KS and 47,XYY were mostly unrecognized but conferred substantially higher risks for metabolic, vascular, and respiratory diseases, which were only partially explained by higher levels of body mass index, deprivation, and smoking.

Rare genetic variants in genes and loci linked to dominant monogenic developmental disorders cause milder related phenotypes in the general population

Many rare monogenic diseases are known to be caused by deleterious variants in thousands of genes, however the same variants can also be found in people without the associated clinical phenotypes. The penetrance of these monogenic variants is generally unknown in the wider population, as they are typically identified in small clinical cohorts of affected individuals and families with highly penetrant variants. Here, we investigated the phenotypic effect of rare, potentially deleterious variants in genes and loci where similar variants are known to cause monogenic developmental disorders (DDs) in a large population cohort. We used UK Biobank to investigate phenotypes associated with rare protein-truncating and missense variants in 599 monoallelic DDG2P genes by using whole-exome-sequencing data from ∼200,000 individuals and rare copy-number variants overlapping known DD loci by using SNP-array data from ∼500,000 individuals. We found that individuals with these likely deleterious variants had a mild DD-related phenotype, including lower fluid intelligence, slower reaction times, lower numeric memory scores, and longer pairs matching times compared to the rest of the UK Biobank cohort. They were also shorter, had a higher BMI, and had significant socioeconomic disadvantages: they were less likely to be employed or be able to work and had a lower income and higher deprivation index. Our findings suggest that many genes routinely tested within pediatric genetics have deleterious variants with intermediate penetrance that may cause lifelong sub-clinical phenotypes in the general adult population.

PLIN1 Haploinsufficiency Causes a Favorable Metabolic Profile

CONTEXT

PLIN1 encodes perilipin-1, which coats lipid droplets in adipocytes and is involved in droplet formation, triglyceride storage, and lipolysis. Rare PLIN1 frameshift variants that extend the translated protein have been described to cause lipodystrophy.

OBJECTIVE

This work aimed to test whether PLIN1 protein-truncating variants (PTVs) cause lipodystrophy in a large population-based cohort.

METHODS

We identified individuals with PLIN1 PTVs in individuals with exome data in the UK Biobank. We performed gene-burden testing for individuals with PLIN1 PTVs. We replicated the associations using data from the T2D Knowledge portal. We performed a phenome-wide association study using publicly available association statistics. A total of 362 791 individuals in the UK Biobank, a population-based cohort, and 43 125 individuals in the T2D Knowledge portal, a type 2 diabetes (T2D) case-control study, were included in the analyses. Main outcome measures included 22 diseases and traits relevant to lipodystrophy.

RESULTS

The 735 individuals with PLIN1 PTVs had a favorable metabolic profile. These individuals had increased high-density lipoprotein cholesterol (0.12 mmol/L; 95% CI, 0.09 to 0.14, P = 2 × 10-18), reduced triglycerides (-0.22 mmol/L; 95% CI, -0.29 to -0.14, P = 3 × 10-11), reduced waist-to-hip ratio (-0.02; 95% CI, -0.02 to -0.01, P = 9 × 10-12), and reduced systolic blood pressure (-1.67 mm Hg; 95% CI, -3.25 to -0.09, P = .05). These associations were consistent in the smaller T2D Knowledge portal cohort. In the UK Biobank, PLIN1 PTVs were associated with reduced risk of myocardial infarction (odds ratio [OR] = 0.59; 95% CI, 0.35 to 0.93, P = .02) and hypertension (OR = 0.85; 95% CI, 0.73 to 0.98, P = .03), but not T2D (OR = 0.99; 95% CI, 0.63-1.51, P = .99).

CONCLUSION

Our study suggests that PLIN1 haploinsufficiency causes a favorable metabolic profile and may protect against cardiovascular disease.

Utility of Diabetes Type-Specific Genetic Risk Scores for the Classification of Diabetes Type Among Multiethnic Youth

OBJECTIVE

Genetic risk scores (GRS) aid classification of diabetes type in White European adult populations. We aimed to assess the utility of GRS in the classification of diabetes type among racially/ethnically diverse youth in the U.S.

RESEARCH DESIGN AND METHODS

We generated type 1 diabetes (T1D)- and type 2 diabetes (T2D)-specific GRS in 2,045 individuals from the SEARCH for Diabetes in Youth study. We assessed the distribution of genetic risk stratified by diabetes autoantibody positive or negative (DAA+/-) and insulin sensitivity (IS) or insulin resistance (IR) and self-reported race/ethnicity (White, Black, Hispanic, and other).

RESULTS

T1D and T2D GRS were strong independent predictors of etiologic type. The T1D GRS was highest in the DAA+/IS group and lowest in the DAA-/IR group, with the inverse relationship observed with the T2D GRS. Discrimination was similar across all racial/ethnic groups but showed differences in score distribution. Clustering by combined genetic risk showed DAA+/IR and DAA-/IS individuals had a greater probability of T1D than T2D. In DAA- individuals, genetic probability of T1D identified individuals most likely to progress to absolute insulin deficiency.

CONCLUSIONS

Diabetes type-specific GRS are consistent predictors of diabetes type across racial/ethnic groups in a U.S. youth cohort, but future work needs to account for differences in GRS distribution by ancestry. T1D and T2D GRS may have particular utility for classification of DAA- children.

Evaluation of Evidence for Pathogenicity Demonstrates That BLK, KLF11, and PAX4 Should Not Be Included in Diagnostic Testing for MODY

Maturity-onset diabetes of the young (MODY) is an autosomal dominant form of monogenic diabetes, reported to be caused by variants in 16 genes. Concern has been raised about whether variants in BLK (MODY11), KLF11 (MODY7), and PAX4 (MODY9) cause MODY. We examined variant-level genetic evidence (cosegregation with diabetes and frequency in population) for published putative pathogenic variants in these genes and used burden testing to test gene-level evidence in a MODY cohort (n = 1,227) compared with a control population (UK Biobank [n = 185,898]). For comparison we analyzed well-established causes of MODY, HNF1A, and HNF4A. The published variants in BLK, KLF11, and PAX4 showed poor cosegregation with diabetes (combined logarithm of the odds [LOD] scores ≤1.2), compared with HNF1A and HNF4A (LOD scores >9), and are all too common to cause MODY (minor allele frequency >4.95 × 10-5). Ultra-rare missense and protein-truncating variants (PTV) were not enriched in a MODY cohort compared with the UK Biobank population (PTV P > 0.05, missense P > 0.1 for all three genes) while HNF1A and HNF4A were enriched (P < 10-6). Findings of sensitivity analyses with different population cohorts supported our results. Variant and gene-level genetic evidence does not support BLK, KLF11, or PAX4 as a cause of MODY. They should not be included in MODY diagnostic genetic testing.

Assessing the Causal Role of Sleep Traits on Glycated Hemoglobin: A Mendelian Randomization Study

OBJECTIVE

To examine the effects of sleep traits on glycated hemoglobin (HbA1c).

RESEARCH DESIGN AND METHODS

This study triangulated evidence across multivariable regression (MVR) and one- (1SMR) and two-sample Mendelian randomization (2SMR) including sensitivity analyses on the effects of five self-reported sleep traits (i.e., insomnia symptoms [difficulty initiating or maintaining sleep], sleep duration, daytime sleepiness, napping, and chronotype) on HbA1c (in SD units) in adults of European ancestry from the UK Biobank (for MVR and 1SMR analyses) (n = 336,999; mean [SD] age 57 [8] years; 54% female) and in the genome-wide association studies from the Meta-Analyses of Glucose and Insulin-Related Traits Consortium (MAGIC) (for 2SMR analysis) (n = 46,368; 53 [11] years; 52% female).

RESULTS

Across MVR, 1SMR, 2SMR, and their sensitivity analyses, we found a higher frequency of insomnia symptoms (usually vs. sometimes or rarely/never) was associated with higher HbA1c (MVR 0.05 SD units [95% CI 0.04-0.06]; 1SMR 0.52 [0.42-0.63]; 2SMR 0.24 [0.11-0.36]). Associations remained, but point estimates were somewhat attenuated after excluding participants with diabetes. For other sleep traits, there was less consistency across methods, with some but not all providing evidence of an effect.

CONCLUSIONS

Our results suggest that frequent insomnia symptoms cause higher HbA1c levels and, by implication, that insomnia has a causal role in type 2 diabetes. These findings could have important implications for developing and evaluating strategies that improve sleep habits to reduce hyperglycemia and prevent diabetes.

A Comparative Safety Analysis of Medicines Based on the UK Pharmacovigilance and General Practice Prescribing Data in England

BACKGROUND/AIM

Adverse drug reactions (ADRs) represent a major concern leading to significant increases in both morbidity and mortality globally. Providing healthcare professionals (HCPs) and patients with real-world data on drug safety is imperative to facilitate informed decision-making. The study aimed to determine the feasibility of creating comparative safety charts for medicines by mapping ADR reporting onto prescribing data.

MATERIALS AND METHODS

Data on serious and fatal ADR reports from the Yellow Card database was mapped onto general practice prescription data in England. The rate of serious and fatal ADR reports per million items prescribed was calculated for commonly-prescribed medicines.

RESULTS

Quantitative comparative analyses for 137 medicines belonging to 26 therapeutic classes were conducted. Significant differences were observed within most therapeutic classes for the rate of serious and fatal ADR reports per prescribing unit.

CONCLUSION

Despite the limitations of ADR reporting and prescribing databases, the study provides a proof-of-concept for the feasibility of mapping ADR reporting onto prescribing data to create comparative safety charts that could support evidence-based decision-making around formulary choices.

Systematic genetic testing for recessively inherited monogenic diabetes: a cross-sectional study in paediatric diabetes clinics

AIMS/HYPOTHESIS

Current clinical guidelines for childhood-onset monogenic diabetes outside infancy are mainly focused on identifying and testing for dominantly inherited, predominantly MODY genes. There are no systematic studies of the recessively inherited causes of monogenic diabetes that are likely to be more common in populations with high rates of consanguinity. We aimed to determine the contribution of recessive causes of monogenic diabetes in paediatric diabetes clinics and to identify clinical criteria by which to select individuals for recessive monogenic diabetes testing.

METHODS

We conducted a cross-sectional study of 1093 children from seven paediatric diabetes clinics across Turkey (a population with high rates of consanguinity). We undertook genetic testing of 50 known dominant and recessive causes of monogenic diabetes for 236 children at low risk of type 1 diabetes. As a comparison, we used monogenic diabetes cases from UK paediatric diabetes clinics (a population with low rates of consanguinity).

RESULTS

Thirty-four children in the Turkish cohort had monogenic diabetes, equating to a minimal prevalence of 3.1%, similar to that in the UK cohort (p = 0.40). Forty-one per cent (14/34) had autosomal recessive causes in contrast to 1.6% (2/122) in the UK monogenic diabetes cohort (p < 0.0001). All conventional criteria for identifying monogenic diabetes (parental diabetes, not requiring insulin treatment, HbA_1c ≤ 58 mmol/mol [≤7.5%] and a composite clinical probability of MODY >10%) assisted the identification of the dominant (all p ≤ 0.0003) but not recessive cases (all p ≥ 0.2) in Turkey. The presence of certain non-autoimmune extra-pancreatic features greatly assisted the identification of recessive (p < 0.0001, OR 66.9) but not dominant cases.

CONCLUSIONS/INTERPRETATION

Recessively inherited mutations are a common cause of monogenic diabetes in populations with high rates of consanguinity. Present MODY-focused genetic testing strategies do not identify affected individuals. To detect all cases of monogenic paediatric diabetes, it is crucial that recessive genes are included in genetic panels and that children are selected for testing if they have certain non-autoimmune extra-pancreatic features in addition to current criteria.

Estimating disease prevalence in large datasets using genetic risk scores

Clinical classification is essential for estimating disease prevalence but is difficult, often requiring complex investigations. The widespread availability of population level genetic data makes novel genetic stratification techniques a highly attractive alternative. We propose a generalizable mathematical framework for determining disease prevalence within a cohort using genetic risk scores. We compare and evaluate methods based on the means of genetic risk scores' distributions; the Earth Mover's Distance between distributions; a linear combination of kernel density estimates of distributions; and an Excess method. We demonstrate the performance of genetic stratification to produce robust prevalence estimates. Specifically, we show that robust estimates of prevalence are still possible even with rarer diseases, smaller cohort sizes and less discriminative genetic risk scores, highlighting the general utility of these approaches. Genetic stratification techniques offer exciting new research tools, enabling unbiased insights into disease prevalence and clinical characteristics unhampered by clinical classification criteria.

Using Mendelian Randomisation methods to understand whether diurnal preference is causally related to mental health

Late diurnal preference has been linked to poorer mental health outcomes, but the understanding of the causal role of diurnal preference on mental health and wellbeing is currently limited. Late diurnal preference is often associated with circadian misalignment (a mismatch between the timing of the endogenous circadian system and behavioural rhythms), so that evening people live more frequently against their internal clock. This study aims to quantify the causal contribution of diurnal preference on mental health outcomes, including anxiety, depression and general wellbeing and test the hypothesis that more misaligned individuals have poorer mental health and wellbeing using an actigraphy-based measure of circadian misalignment. Multiple Mendelian Randomisation (MR) approaches were used to test causal pathways between diurnal preference and seven well-validated mental health and wellbeing outcomes in up to 451,025 individuals. In addition, observational analyses tested the association between a novel, objective measure of behavioural misalignment (Composite Phase Deviation, CPD) and seven mental health and wellbeing outcomes. Using genetic instruments identified in the largest GWAS for diurnal preference, we provide robust evidence that early diurnal preference is protective for depression and improves wellbeing. For example, using one-sample MR, a twofold higher genetic liability of morningness was associated with lower odds of depressive symptoms (OR: 0.92, 95% CI: 0.88, 0.97). It is possible that behavioural factors including circadian misalignment may contribute in the chronotype depression relationship, but further work is needed to confirm these findings.

DR15-DQ6 remains dominantly protective against type 1 diabetes throughout the first five decades of life

AIMS/HYPOTHESIS

Among white European children developing type 1 diabetes, the otherwise common HLA haplotype DR15-DQ6 is rare, and highly protective. Adult-onset type 1 diabetes is now known to represent more overall cases than childhood onset, but it is not known whether DR15-DQ6 is protective in older-adult-onset type 1 diabetes. We sought to quantify DR15-DQ6 protection against type 1 diabetes as age of onset increased.

METHODS

In two independent cohorts we assessed the proportion of type 1 diabetes cases presenting through the first 50 years of life with DR15-DQ6, compared with population controls. In the After Diabetes Diagnosis Research Support System-2 (ADDRESS-2) cohort (n = 1458) clinician-diagnosed type 1 diabetes was confirmed by positivity for one or more islet-specific autoantibodies. In UK Biobank (n = 2502), we estimated type 1 diabetes incidence rates relative to baseline HLA risk for each HLA group using Poisson regression. Analyses were restricted to white Europeans and were performed in three groups according to age at type 1 diabetes onset: 0-18 years, 19-30 years and 31-50 years.

RESULTS

DR15-DQ6 was protective against type 1 diabetes through to age 50 years (OR < 1 for each age group, all p < 0.001). The following ORs for type 1 diabetes, relative to a neutral HLA genotype, were observed in ADDRESS-2: age 5-18 years OR 0.16 (95% CI 0.08, 0.31); age 19-30 years OR 0.10 (0.04, 0.23); and age 31-50 years OR 0.37 (0.21, 0.68). DR15-DQ6 also remained highly protective at all ages in UK Biobank. Without DR15-DQ6, the presence of major type 1 diabetes high-risk haplotype (either DR3-DQ2 or DR4-DQ8) was associated with increased risk of type 1 diabetes.

CONCLUSIONS/INTERPRETATION

HLA DR15-DQ6 confers dominant protection from type 1 diabetes across the first five decades of life.

Is disrupted sleep a risk factor for Alzheimer's disease? Evidence from a two-sample Mendelian randomization analysis

Background

It is established that Alzheimer’s disease (AD) patients experience sleep disruption. However, it remains unknown whether disruption in the quantity, quality or timing of sleep is a risk factor for the onset of AD.

Methods

We used the largest published genome-wide association studies of self-reported and accelerometer-measured sleep traits (chronotype, duration, fragmentation, insomnia, daytime napping and daytime sleepiness), and AD. Mendelian randomization (MR) was used to estimate the causal effect of self-reported and accelerometer-measured sleep parameters on AD risk.

Results

Overall, there was little evidence to support a causal effect of sleep traits on AD risk. There was some suggestive evidence that self-reported daytime napping was associated with lower AD risk [odds ratio (OR): 0.70, 95% confidence interval (CI): 0.50–0.99). Some other sleep traits (accelerometer-measured ‘eveningness’ and sleep duration, and self-reported daytime sleepiness) had ORs of a similar magnitude to daytime napping, but were less precisely estimated.

Conclusions

Overall, we found very limited evidence to support a causal effect of sleep traits on AD risk. Our findings provide tentative evidence that daytime napping may reduce AD risk. Given that this is the first MR study of multiple self-report and objective sleep traits on AD risk, findings should be replicated using independent samples when such data become available.

A genome-wide association study identifies 5 loci associated with frozen shoulder and implicates diabetes as a causal risk factor

Frozen shoulder is a painful condition that often requires surgery and affects up to 5% of individuals aged 40-60 years. Little is known about the causes of the condition, but diabetes is a strong risk factor. To begin to understand the biological mechanisms involved, we aimed to identify genetic variants associated with frozen shoulder and to use Mendelian randomization to test the causal role of diabetes. We performed a genome-wide association study (GWAS) of frozen shoulder in the UK Biobank using data from 10,104 cases identified from inpatient, surgical and primary care codes. We used data from FinnGen for replication and meta-analysis. We used one-sample and two-sample Mendelian randomization approaches to test for a causal association of diabetes with frozen shoulder. We identified five genome-wide significant loci. The most significant locus (lead SNP rs28971325; OR = 1.20, [95% CI: 1.16-1.24], p = 5x10-29) contained WNT7B. This variant was also associated with Dupuytren's disease (OR = 2.31 [2.24, 2.39], p<1x10-300) as were a further two of the frozen shoulder associated variants. The Mendelian randomization results provided evidence that type 1 diabetes is a causal risk factor for frozen shoulder (OR = 1.03 [1.02-1.05], p = 3x10-6). There was no evidence that obesity was causally associated with frozen shoulder, suggesting that diabetes influences risk of the condition through glycemic rather than mechanical effects. We have identified genetic loci associated with frozen shoulder. There is a large overlap with Dupuytren's disease associated loci. Diabetes is a likely causal risk factor. Our results provide evidence of biological mechanisms involved in this common painful condition.

Genetic evidence that higher central adiposity causes gastro-oesophageal reflux disease: a Mendelian randomization study

Background

Gastro-oesophageal reflux disease (GORD) is associated with multiple risk factors but determining causality is difficult. We used a genetic approach [Mendelian randomization (MR)] to identify potential causal modifiable risk factors for GORD.

Methods

We used data from 451 097 European participants in the UK Biobank and defined GORD using hospital-defined ICD10 and OPCS4 codes and self-report data (N = 41 024 GORD cases). We tested observational and MR-based associations between GORD and four adiposity measures [body mass index (BMI), waist–hip ratio (WHR), a metabolically favourable higher body-fat percentage and waist circumference], smoking status, smoking frequency and caffeine consumption.

Results

Observationally, all adiposity measures were associated with higher odds of GORD. Ever and current smoking were associated with higher odds of GORD. Coffee consumption was associated with lower odds of GORD but, among coffee drinkers, more caffeinated-coffee consumption was associated with higher odds of GORD. Using MR, we provide strong evidence that higher WHR and higher WHR adjusted for BMI lead to GORD. There was weak evidence that higher BMI, body-fat percentage, coffee drinking or smoking caused GORD, but only the observational effects for BMI and body-fat percentage could be excluded. This MR estimated effect for WHR equates to a 1.23-fold higher odds of GORD per 5-cm increase in waist circumference.

Conclusions

These results provide strong evidence that a higher waist–hip ratio leads to GORD. Our study suggests that central fat distribution is crucial in causing GORD rather than overall weight.

Genetic determinants of daytime napping and effects on cardiometabolic health

Daytime napping is a common, heritable behavior, but its genetic basis and causal relationship with cardiometabolic health remain unclear. Here, we perform a genome-wide association study of self-reported daytime napping in the UK Biobank (n = 452,633) and identify 123 loci of which 61 replicate in the 23andMe research cohort (n = 541,333). Findings include missense variants in established drug targets for sleep disorders (HCRTR1, HCRTR2), genes with roles in arousal (TRPC6, PNOC), and genes suggesting an obesity-hypersomnolence pathway (PNOC, PATJ). Association signals are concordant with accelerometer-measured daytime inactivity duration and 33 loci colocalize with loci for other sleep phenotypes. Cluster analysis identifies three distinct clusters of nap-promoting mechanisms with heterogeneous associations with cardiometabolic outcomes. Mendelian randomization shows potential causal links between more frequent daytime napping and higher blood pressure and waist circumference.

Sleep classification from wrist-worn accelerometer data using random forests

Accurate and low-cost sleep measurement tools are needed in both clinical and epidemiological research. To this end, wearable accelerometers are widely used as they are both low in price and provide reasonably accurate estimates of movement. Techniques to classify sleep from the high-resolution accelerometer data primarily rely on heuristic algorithms. In this paper, we explore the potential of detecting sleep using Random forests. Models were trained using data from three different studies where 134 adult participants (70 with sleep disorder and 64 good healthy sleepers) wore an accelerometer on their wrist during a one-night polysomnography recording in the clinic. The Random forests were able to distinguish sleep-wake states with an F1 score of 73.93% on a previously unseen test set of 24 participants. Detecting when the accelerometer is not worn was also successful using machine learning ([Formula: see text]), and when combined with our sleep detection models on day-time data provide a sleep estimate that is correlated with self-reported habitual nap behaviour ([Formula: see text]). These Random forest models have been made open-source to aid further research. In line with literature, sleep stage classification turned out to be difficult using only accelerometer data.

Clinical Features and Genetic Risk of Demyelination Following Anti-TNF Treatment

BACKGROUND

Anti-TNF exposure has been linked to demyelination events. We sought to describe the clinical features of demyelination events following anti-TNF treatment and to test whether affected patients were genetically predisposed to multiple sclerosis [MS].

METHODS

We conducted a case-control study to describe the clinical features of demyelination events following anti-TNF exposure. We compared genetic risk scores [GRS], calculated using carriage of 43 susceptibility loci for MS, in 48 cases with 1219 patients exposed to anti-TNF who did not develop demyelination.

RESULTS

Overall, 39 [74%] cases were female. The median age [range] of patients at time of demyelination was 41.5 years [20.7-63.2]. The median duration of anti-TNF treatment was 21.3 months [0.5-99.4] and 19 [36%] patients were receiving concomitant immunomodulators. Most patients had central demyelination affecting the brain, spinal cord, or both. Complete recovery was reported in 12 [23%] patients after a median time of 6.8 months [0.1-28.7]. After 33.0 months of follow-up, partial recovery was observed in 29 [55%] patients, relapsing and remitting episodes in nine [17%], progressive symptoms in three [6%]: two [4%] patients were diagnosed with MS. There was no significant difference between MS GRS scores in cases (mean -3.5 × 10-4, standard deviation [SD] 0.0039) and controls [mean -1.1 × 10-3, SD 0.0042] [p = 0.23].

CONCLUSIONS

Patients who experienced demyelination events following anti-TNF exposure were more likely female, less frequently treated with an immunomodulator, and had a similar genetic risk to anti-TNF exposed controls who did not experience demyelination events. Large prospective studies with pre-treatment neuroimaging are required to identify genetic susceptibility loci.