Zoish: A Novel Feature Selection Approach Leveraging Shapley Additive Values for Machine Learning Applications in Healthcare.

In the intricate landscape of healthcare analytics, effective feature selection is a prerequisite for generating robust predictive models, especially given the common challenges of sample sizes and potential biases. Zoish uniquely addresses these issues by employing Shapley additive values-an idea rooted in cooperative game theory-to enable both transparent and automated feature selection. Unlike existing tools, Zoish is versatile, designed to seamlessly integrate with an array of machine learning libraries including scikit-learn, XGBoost, CatBoost, and imbalanced-learn.The distinct advantage of Zoish lies in its dual algorithmic approach for calculating Shapley values, allowing it to efficiently manage both large and small datasets. This adaptability renders it exceptionally suitable for a wide spectrum of healthcare-related tasks. The tool also places a strong emphasis on interpretability, providing comprehensive visualizations for analyzed features. Its customizable settings offer users fine-grained control over feature selection, thus optimizing for specific predictive objectives.This manuscript elucidates the mathematical framework underpinning Zoish and how it uniquely combines local and global feature selection into a single, streamlined process. To validate Zoish's efficiency and adaptability, we present case studies in breast cancer prediction and Montreal Cognitive Assessment (MoCA) prediction in Parkinson's disease, along with evaluations on 300 synthetic datasets. These applications underscore Zoish's unparalleled performance in diverse healthcare contexts and against its counterparts.

Dose response of running on blood biomarkers of wellness in generally healthy individuals.

Exercise is effective toward delaying or preventing chronic disease, with a large body of evidence supporting its effectiveness. However, less is known about the specific healthspan-promoting effects of exercise on blood biomarkers in the disease-free population. In this work, we examine 23,237 generally healthy individuals who self-report varying weekly running volumes and compare them to 4,428 generally healthy sedentary individuals, as well as 82 professional endurance runners. We estimate the significance of differences among blood biomarkers for groups of increasing running levels using analysis of variance (ANOVA), adjusting for age, gender, and BMI. We attempt and add insight to our observational dataset analysis via two-sample Mendelian randomization (2S-MR) using large independent datasets. We find that self-reported running volume associates with biomarker signatures of improved wellness, with some serum markers apparently being principally modified by BMI, whereas others show a dose-effect with respect to running volume. We further detect hints of sexually dimorphic serum responses in oxygen transport and hormonal traits, and we also observe a tendency toward pronounced modifications in magnesium status in professional endurance athletes. Thus, our results further characterize blood biomarkers of exercise and metabolic health, particularly regarding dose-effect relationships, and better inform personalized advice for training and performance.

Polygenic Risk Score-Based Association Analysis of Speech-in-Noise and Hearing Threshold Measures in Healthy Young Adults with Self-reported Normal Hearing.

PURPOSE: Speech-in-noise (SIN) traits exhibit high inter-subject variability, even for healthy young adults reporting normal hearing. Emerging evidence suggests that genetic variability could influence inter-subject variability in SIN traits. Genome-wide association studies (GWAS) have uncovered the polygenic architecture of various adult-onset complex human conditions. Polygenic risk scores (PRS) summarize complex genetic susceptibility to quantify the degree of genetic risk for health conditions. The present study conducted PRS-based association analyses to identify PRS risk factors for SIN and hearing threshold measures in 255 healthy young adults (18-40 years) with self-reported normal hearing. METHODS: Self-reported SIN perception abilities were assessed by the Speech, Spatial, and Qualities of Hearing Scale (SSQ12). QuickSIN and audiometry (0.25-16 kHz) were performed on 218 participants. Saliva-derived DNA was used for low-pass whole genome sequencing, and 2620 PRS variables for various traits were calculated using the models derived from the polygenic risk score (PGS) catalog. The regression analysis was conducted to identify predictors for SSQ12, QuickSIN, and better ear puretone averages at conventional (PTA0.5-2), high (PTA4-8), and extended-high (PTA12.5-16) frequency ranges. RESULTS: Participants with a higher genetic predisposition to HDL cholesterol reported better SSQ12. Participants with high PRS to dementia revealed significantly elevated PTA4-8, and those with high PRS to atrial fibrillation and flutter revealed significantly elevated PTA12.5-16. CONCLUSION: These results indicate that healthy individuals with polygenic risk of certain health conditions could exhibit a subclinical decline in hearing health measures at young ages, decades before clinically meaningful SIN deficits and hearing loss could be observed. PRS could be used to identify high-risk individuals to prevent hearing health conditions by promoting a healthy lifestyle.

The promise of digital healthcare technologies.

Digital health technologies have been in use for many years in a wide spectrum of healthcare scenarios. This narrative review outlines the current use and the future strategies and significance of digital health technologies in modern healthcare applications. It covers the current state of the scientific field (delineating major strengths, limitations, and applications) and envisions the future impact of relevant emerging key technologies. Furthermore, we attempt to provide recommendations for innovative approaches that would accelerate and benefit the research, translation and utilization of digital health technologies.

Identifying Health-Related Conditions Associated with Tinnitus in Young Adults.

OBJECTIVE: The present study investigated the epidemic of tinnitus in college-aged young adults. Our first objective was to identify health conditions associated with tinnitus in young adults. The second objective was to evaluate the predictive utility of some known risk factors. STUDY DESIGN: A cross-sectional design was used to investigate the prevalence and risk factors for tinnitus. SETTING: A questionnaire was distributed, reaching out to a large college-aged population. A total of 2258 young adults aged 18-30 years were recruited from April 2021 to February 2022. INTERVENTIONS: A questionnaire was administered to investigate the epidemiology of tinnitus in a population of college-aged young adults. RESULTS: About 17.7% of young adults reported bothersome tinnitus perception lasting for ≥5 min in the last 12 months. The prevalence of chronic tinnitus (bothersome tinnitus for ≥1 year) and acute tinnitus (bothersome tinnitus for <1 year) was 10.6% and 7.1%, respectively. About 19% of the study sample reported at least one health condition. Individuals reporting head injury, hypertension, heart disease, scarlet fever, and malaria showed significantly higher odds of reporting chronic tinnitus. Meningitis and self-reported hearing loss showed significant associations with bothersome tinnitus. The prevalence of chronic tinnitus was significantly higher in males reporting high noise exposure, a positive history of reoccurring ear infections, European ethnic background, and a positive health history. Risk modeling showed that noise exposure was the most important risk factor for chronic tinnitus, followed by sex, reoccurring ear infections, and a history of any health condition. A positive history of COVID-19 and self-reported severity showed no association with tinnitus. Individuals reporting reoccurring ear infections showed a significantly higher prevalence of COVID-19. CONCLUSIONS: While young adults with health conditions are at a higher risk of reporting tinnitus, the predictive utility of a positive health history remains relatively low, possibly due to weak associations between health conditions and tinnitus. Noise, male sex, reoccurring ear infections, European ethnicity, and a positive health history revealed higher odds of reporting chronic tinnitus than their counterparts. These risk factors collectively explained about 16% variability in chronic tinnitus, which highlights the need for identifying other risk factors for chronic tinnitus in young adults.

Genome-wide structural variant analysis identifies risk loci for non-Alzheimer's dementias.

We characterized the role of structural variants, a largely unexplored type of genetic variation, in two non-Alzheimer's dementias, namely Lewy body dementia (LBD) and frontotemporal dementia (FTD)/amyotrophic lateral sclerosis (ALS). To do this, we applied an advanced structural variant calling pipeline (GATK-SV) to short-read whole-genome sequence data from 5,213 European-ancestry cases and 4,132 controls. We discovered, replicated, and validated a deletion in TPCN1 as a novel risk locus for LBD and detected the known structural variants at the C9orf72 and MAPT loci as associated with FTD/ALS. We also identified rare pathogenic structural variants in both LBD and FTD/ALS. Finally, we assembled a catalog of structural variants that can be mined for new insights into the pathogenesis of these understudied forms of dementia.

Large-scale rare variant burden testing in Parkinson's disease.

Parkinson's disease has a large heritable component and genome-wide association studies have identified over 90 variants with disease-associated common variants, providing deeper insights into the disease biology. However, there have not been large-scale rare variant analyses for Parkinson's disease. To address this gap, we investigated the rare genetic component of Parkinson's disease at minor allele frequencies <1%, using whole genome and whole exome sequencing data from 7184 Parkinson's disease cases, 6701 proxy cases and 51 650 healthy controls from the Accelerating Medicines Partnership Parkinson's disease (AMP-PD) initiative, the National Institutes of Health, the UK Biobank and Genentech. We performed burden tests meta-analyses on small indels and single nucleotide protein-altering variants, prioritized based on their predicted functional impact. Our work identified several genes reaching exome-wide significance. Two of these genes, GBA1 and LRRK2, have variants that have been previously implicated as risk factors for Parkinson's disease, with some variants in LRRK2 resulting in monogenic forms of the disease. We identify potential novel risk associations for variants in B3GNT3, AUNIP, ADH5, TUBA1B, OR1G1, CAPN10 and TREML1 but were unable to replicate the observed associations across independent datasets. Of these, B3GNT3 and TREML1 could provide new evidence for the role of neuroinflammation in Parkinson's disease. To date, this is the largest analysis of rare genetic variants in Parkinson's disease.

Genome-Wide Analysis of Structural Variants in Parkinson Disease.

OBJECTIVE: Identification of genetic risk factors for Parkinson disease (PD) has to date been primarily limited to the study of single nucleotide variants, which only represent a small fraction of the genetic variation in the human genome. Consequently, causal variants for most PD risk are not known. Here we focused on structural variants (SVs), which represent a major source of genetic variation in the human genome. We aimed to discover SVs associated with PD risk by performing the first large-scale characterization of SVs in PD. METHODS: We leveraged a recently developed computational pipeline to detect and genotype SVs from 7,772 Illumina short-read whole genome sequencing samples. Using this set of SV variants, we performed a genome-wide association study using 2,585 cases and 2,779 controls and identified SVs associated with PD risk. Furthermore, to validate the presence of these variants, we generated a subset of matched whole-genome long-read sequencing data. RESULTS: We genotyped and tested 3,154 common SVs, representing over 412 million nucleotides of previously uncatalogued genetic variation. Using long-read sequencing data, we validated the presence of three novel deletion SVs that are associated with risk of PD from our initial association analysis, including a 2 kb intronic deletion within the gene LRRN4. INTERPRETATION: We identified three SVs associated with genetic risk of PD. This study represents the most comprehensive assessment of the contribution of SVs to the genetic risk of PD to date. ANN NEUROL 2023;93:1012-1022.

HIGH-PERFORMANCE COMPUTING MEETS HIGH-PERFORMANCE MEDICINE.

The following sections are included: Introduction, Background, and Motivation, Workshop Presenters, References.

Meta-Prediction of Coronary Artery Disease Risk.

Coronary artery disease (CAD) remains the leading cause of mortality and morbidity worldwide. Recent advances in large-scale genome-wide association studies have highlighted the potential of genetic risk, captured as polygenic risk scores (PRS), in clinical prevention. However, the current clinical utility of PRS models is limited to identifying high-risk populations based on the top percentiles of genetic susceptibility. While some studies have attempted integrative prediction using genetic and non-genetic factors, many of these studies have been cross-sectional and focused solely on risk stratification. Our primary objective in this study was to integrate unmodifiable (age / genetics) and modifiable (clinical / biometric) risk factors into a prospective prediction framework which also produces actionable and personalized risk estimates for the purpose of CAD prevention in a heterogenous adult population. Thus, we present an integrative, omnigenic, meta-prediction framework that effectively captures CAD risk subgroups, primarily distinguished by degree and nature of genetic risk, with distinct risk reduction profiles predicted from standard clinical interventions. Initial model development considered ~ 2,000 predictive features, including demographic data, lifestyle factors, physical measurements, laboratory tests, medication usage, diagnoses, and genetics. To power our meta-prediction approach, we stratified the UK Biobank into two primary cohorts: 1) a prevalent CAD cohort used to train baseline and prospective predictive models for contributing risk factors and diagnoses, and 2) an incident CAD cohort used to train the final CAD incident risk prediction model. The resultant 10-year incident CAD risk model is composed of 35 derived meta-features from models trained on the prevalent risk cohort, most of which are predicted baseline diagnoses with multiple embedded PRSs. This model achieved an AUC of 0.81 and macro-averaged F1-score of 0.65, outperforming standard clinical scores and prior integrative models. We further demonstrate that individualized risk reduction profiles can be derived from this model, with genetic risk mediating the degree of risk reduction achieved by standard clinical interventions.

A genome-wide association study of tinnitus reveals shared genetic links to neuropsychiatric disorders.

Tinnitus, a phantom perception of sound in the absence of any external sound source, is a prevalent health condition often accompanied by psychiatric comorbidities. Recent genome-wide association studies (GWAS) highlighted a polygenic nature of tinnitus susceptibility. A shared genetic component between tinnitus and psychiatric conditions remains elusive. Here we present a GWAS using the UK Biobank to investigate the genetic processes linked to tinnitus and tinnitus-related distress, followed by gene-set enrichment analyses. The UK Biobank sample comprised 132,438 individuals with tinnitus and genotype data. Among the study sample, 38,525 individuals reported tinnitus, and 26,889 participants mentioned they experienced tinnitus-related distress in daily living. The genome-wide association analyses were conducted on tinnitus and tinnitus-related distress. We conducted enrichment analyses using FUMA to further understand the genetic processes linked to tinnitus and tinnitus-related distress. A genome-wide significant locus (lead SNP: rs71595470) for tinnitus was obtained in the vicinity of GPM6A. Nineteen independent loci reached suggestive association with tinnitus. Fifteen independent loci reached suggestive association with tinnitus-related distress. The enrichment analysis revealed a shared genetic component between tinnitus and psychiatric traits, such as bipolar disorder, feeling worried, cognitive ability, fast beta electroencephalogram, and sensation seeking. Metabolic, cardiovascular, hematological, and pharmacological gene sets revealed a significant association with tinnitus. Anxiety and stress-related gene sets revealed a significant association with tinnitus-related distress. The GWAS signals for tinnitus were enriched in the hippocampus and cortex, and for tinnitus-related distress were enriched in the brain and spinal cord. This study provides novel insights into genetic processes associated with tinnitus and tinnitus-related distress and demonstrates a shared genetic component underlying tinnitus and psychiatric conditions. Further collaborative attempts are necessary to identify genetic components underlying the phenotypic heterogeneity in tinnitus and provide biological insight into the etiology.

Machine learning-based prediction of cognitive outcomes in de novo Parkinson's disease.

Cognitive impairment is a debilitating symptom in Parkinson's disease (PD). We aimed to establish an accurate multivariate machine learning (ML) model to predict cognitive outcome in newly diagnosed PD cases from the Parkinson's Progression Markers Initiative (PPMI). Annual cognitive assessments over an 8-year time span were used to define two cognitive outcomes of (i) cognitive impairment, and (ii) dementia conversion. Selected baseline variables were organized into three subsets of clinical, biofluid and genetic/epigenetic measures and tested using four different ML algorithms. Irrespective of the ML algorithm used, the models consisting of the clinical variables performed best and showed better prediction of cognitive impairment outcome over dementia conversion. We observed a marginal improvement in the prediction performance when clinical, biofluid, and epigenetic/genetic variables were all included in one model. Several cerebrospinal fluid measures and an epigenetic marker showed high predictive weighting in multiple models when included alongside clinical variables.

Decoding the transcriptome of calcified atherosclerotic plaque at single-cell resolution.

Atherogenesis involves an interplay of inflammation, tissue remodeling and cellular transdifferentiation (CTD), making it especially difficult to precisely delineate its pathophysiology. Here we use single-cell RNA sequencing and systems-biology approaches to analyze the transcriptional profiles of vascular smooth muscle cells (VSMCs) and endothelial cells (ECs) in calcified atherosclerotic core (AC) plaques and patient-matched proximal adjacent (PA) portions of carotid artery tissue from patients undergoing carotid endarterectomy. Our results reveal an anatomic distinction whereby PA cells express inflammatory mediators, while cells expressing matrix-secreting genes occupy a majority of the AC region. Systems biology analysis indicates that inflammation in PA ECs and VSMCs may be driven by TNFa signaling. Furthermore, we identify POSTN, SPP1 and IBSP in AC VSMCs, and ITLN1, SCX and S100A4 in AC ECs as possible candidate drivers of CTD in the atherosclerotic core. These results establish an anatomic framework for atherogenesis which forms the basis for exploration of a site-specific strategy for disruption of disease progression.

Suprathreshold Auditory Measures for Detecting Early-Stage Noise-Induced Hearing Loss in Young Adults.

BACKGROUND: Over 1 billion young adults are at risk for developing noise-induced hearing loss (NIHL) due to their habit of listening to music at loud levels. The gold standard for detecting NIHL is the audiometric notch around 3,000 to 6,000 Hz observed in pure tone audiogram. However, recent studies suggested that suprathreshold auditory measures might be more sensitive to detect early-stage NIHL in young adults. PURPOSE: The present study compared suprathreshold measures in individuals with high and low noise exposure backgrounds (NEBs). We hypothesized that individuals with high NEB would exhibit reduced performance on suprathreshold measures than those with low NEB. STUDY SAMPLE: An initial sample of 100 English-speaking healthy adults (18-35 years; females = 70) was obtained from five university classes. We identified 15 participants with the lowest NEB scores (10 females) and 15 participants with the highest NEB scores (10 females). We selected a sample of healthy young adults with no history of middle ear infection, and those in the low NEB group were selected with no history of impulse noise exposure. DATA COLLECTION AND ANALYSIS: The study included conventional audiometry, extended high-frequency audiometry, middle ear muscle reflex (MEMR) thresholds, distortion-product otoacoustic emissions (DPOAEs), QuickSIN, and suprathreshold auditory brainstem response (ABR) measures. We used independent sample t-tests, correlation coefficients, and linear mixed model analysis to compare the audiometric measures between the NEB groups. RESULTS: The prevalence of audiometric notch was low in the study sample, even for individuals with high NEB. We found that: (1) individuals with high NEB revealed significantly reduced QuickSIN performance than those with low NEB; (2) music exposure via earphone revealed a significant association with QuickSIN; (3) individuals with high NEB revealed significantly reduced DPOAEs and ABR wave I amplitude compared with individuals with low NEB; (4) MEMR and ABR latency measures showed a modest association with NEB; and (5) audiometric thresholds across the frequency range did not show statistically significant association with NEB. CONCLUSION: Our results suggest that young adults with high NEB might exhibit impaired peripheral neural coding deficits leading to reduced speech-in-noise (SIN) performance despite clinically normal hearing thresholds. SIN measures might be more sensitive than audiometric notch for detecting early-stage NIHL in young adults.

Nidogen-1 could play a role in diabetic kidney disease development in type 2 diabetes: a genome-wide association meta-analysis.

BACKGROUND: Diabetic kidney disease (DKD) affects about 40% of patients with diabetes. It is incurable and usually leads to end-stage renal disease (ESRD). The pathogenesis of DKD is still not fully understood, and the genetics of DKD have not yet been extensively studied. In this study, we investigate the genetic basis of DKD in type 2 diabetes (T2D) to provide more insights into the pathogenesis of the disease. RESULTS: Using the data provided by the UK Biobank (UKBB), we performed a DKD genome-wide association study (GWAS) in 13,123 individuals with T2D as well as two creatinine estimated glomerular filtration rate (eGFR) GWA studies: one in 26,786 individuals with T2D and the other in 339,080 non-diabetic individuals. We also conducted a DKD GWAS meta-analysis combining our results with those published by the surrogate markers for micro- and macro-vascular hard endpoints for Innovative diabetes Tools (SUMMIT) consortium. We confirm two loci previously reported to be associated with chronic kidney disease (CKD) and eGFR in T2D. The UMOD-PDILT locus is associated with DKD (P = 1.17E-09) as well as creatinine eGFR in both people with T2D (P = 1.31E-15) and people without diabetes (P = 3.95E-73). The PRKAG2 locus is associated with creatinine eGFR in people with (P = 2.78E-10) and without (P = 5.65E-72) T2D. Our meta-analysis reveals a novel association between DKD and variant rs72763500 (chr1:236116561) which is a splicing quantitative trait locus (sQTL) for nidogen-1 (NID1) gene. CONCLUSION: Our data confirm two loci previously reported in association with CKD and creatinine eGFR in T2D. It also suggests that NID1, a major component of the renal tubular basement membrane, could play a role in DKD development in T2D. While our NID1 finding remains to be replicated, it is a step toward a more comprehensive understanding of DKD pathogenesis.

Genetically-informed prediction of short-term Parkinson's disease progression.

Parkinson's disease (PD) treatments modify disease symptoms but have not been shown to slow progression, characterized by gradual and varied motor and non-motor changes overtime. Variation in PD progression hampers clinical research, resulting in long and expensive clinical trials prone to failure. Development of models for short-term PD progression prediction could be useful for shortening the time required to detect disease-modifying drug effects in clinical studies. PD progressors were defined by an increase in MDS-UPDRS scores at 12-, 24-, and 36-months post-baseline. Using only baseline features, PD progression was separately predicted across all timepoints and MDS-UPDRS subparts in independent, optimized, XGBoost models. These predictions plus baseline features were combined into a meta-predictor for 12-month MDS UPDRS Total progression. Data from the Parkinson's Progression Markers Initiative (PPMI) were used for training with independent testing on the Parkinson's Disease Biomarkers Program (PDBP) cohort. 12-month PD total progression was predicted with an F-measure 0.77, ROC AUC of 0.77, and PR AUC of 0.76 when tested on a hold-out PPMI set. When tested on PDBP we achieve a F-measure 0.75, ROC AUC of 0.74, and PR AUC of 0.73. Exclusion of genetic predictors led to the greatest loss in predictive accuracy; ROC AUC of 0.66, PR AUC of 0.66-0.68 for both PPMI and PDBP testing. Short-term PD progression can be predicted with a combination of survey-based, neuroimaging, physician examination, and genetic predictors. Dissection of the interplay between genetic risk, motor symptoms, non-motor symptoms, and longer-term expected rates of progression enable generalizable predictions.

Rapid, Reference-Free human genotype imputation with denoising autoencoders.

Genotype imputation is a foundational tool for population genetics. Standard statistical imputation approaches rely on the co-location of large whole-genome sequencing-based reference panels, powerful computing environments, and potentially sensitive genetic study data. This results in computational resource and privacy-risk barriers to access to cutting-edge imputation techniques. Moreover, the accuracy of current statistical approaches is known to degrade in regions of low and complex linkage disequilibrium. Artificial neural network-based imputation approaches may overcome these limitations by encoding complex genotype relationships in easily portable inference models. Here, we demonstrate an autoencoder-based approach for genotype imputation, using a large, commonly used reference panel, and spanning the entirety of human chromosome 22. Our autoencoder-based genotype imputation strategy achieved superior imputation accuracy across the allele-frequency spectrum and across genomes of diverse ancestry, while delivering at least fourfold faster inference run time relative to standard imputation tools.

Gain of function of a metalloproteinase associated with multiple myeloma, bicuspid aortic valve, and Von Hippel-Lindau syndrome.

A patient diagnosed with multiple myeloma, bicuspid aortic valve, and Von Hippel-Lindau syndrome underwent whole-exome sequencing seeking a unified genetic cause for these three pathologies. The patient possessed a single-point mutation of arginine to cysteine (R24C) in the N-terminal region(pro-domain) of matrix metalloproteinase 9 (MMP-9). The pro-domain interacts with the catalytic site of this enzyme rendering it inactive. MMP-9 has previously been associated with all three pathologies suffered by the patient. We hypothesized that the observed mutation in the pro-domain would influence the activity of this enzyme. We expressed recombinant versions of MMP-9 and an investigation of their biochemical properties revealed that MMP-9 R24C is a constitutively active zymogen. To our knowledge, this is the first example of a mutation that discloses catalytic activity in the pro-form in any of the 24 human MMPs.

Endocannabinoid dysfunction in neurological disease: neuro-ocular DAGLA-related syndrome.

The endocannabinoid system is a highly conserved and ubiquitous signalling pathway with broad-ranging effects. Despite critical pathway functions, gene variants have not previously been conclusively linked to human disease. We identified nine children from eight families with heterozygous, de novo truncating variants in the last exon of DAGLA with a neuro-ocular phenotype characterized by developmental delay, ataxia and complex oculomotor abnormality. All children displayed paroxysms of nystagmus or eye deviation accompanied by compensatory head posture and worsened incoordination most frequently after waking. RNA sequencing showed clear expression of the truncated transcript and no differences were found between mutant and wild-type DAGLA activity. Immunofluorescence staining of patient-derived fibroblasts and HEK cells expressing the mutant protein showed distinct perinuclear aggregation not detected in control samples. This report establishes truncating variants in the last DAGLA exon as the cause of a unique paediatric syndrome. Because enzymatic activity was preserved, the observed mislocalization of the truncated protein may account for the observed phenotype. Potential mechanisms include DAGLA haploinsufficiency at the plasma membrane or dominant negative effect. To our knowledge, this is the first report directly linking an endocannabinoid system component with human genetic disease and sets the stage for potential future therapeutic avenues.