PheWAS: demonstrating the feasibility of a phenome-wide scan to discover gene-disease associations

A comprehensive review of ICU readmission prediction models: From statistical methods to deep learning approaches

The prediction of Intensive Care Unit (ICU) readmission has become a crucial area of research due to the increasing demand for ICU resources and the need to provide timely interventions to critically ill patients. In recent years, several studies have explored the use of statistical, machine learning (ML), and deep learning (DL) models to predict ICU readmission. This review paper presents an extensive overview of these studies and discusses the challenges associated with ICU readmission prediction. We categorize the studies based on the type of model used and evaluate their strengths and limitations. We also discuss the performance metrics used to evaluate the models and their potential clinical applications. In addition, this review explores current methodologies, data usage, and recent advances in interpretability and explainable AI for medical applications, offering insights to guide future research and development in this field. Finally, we identify gaps in the current literature and provide recommendations for future research. Recent advances like ML and DL have moderately improved the prediction of the risk of ICU readmission. However, more progress is needed to reach the precision required to build computerized decision support tools.

Beyond Phecodes: leveraging PheMAP to identify patients lacking diagnosis codes in electronic health records

OBJECTIVE

Diagnosis codes documented in electronic health records (EHR) are often relied upon to clinically phenotype patients for biomedical research. However, these diagnoses can be incomplete and inaccurate, leading to false negatives when searching for patients with phenotypes of interest. This study aims to determine whether PheMAP, a comprehensive knowledgebase integrating multiple clinical terminologies beyond diagnosis to capture phenotypes, can effectively identify patients lacking relevant EHR diagnosis codes.

MATERIALS AND METHODS

We investigated a collection of 3.5 million patient records from Vanderbilt University Medical Center's EHR and focused on 4 well-studied phenotypes: (1) type 2 diabetes mellitus (T2DM), (2) dementia, (3) prostate cancer, and (4) sensorineural hearing loss. We applied PheMAP to match structured concepts in patient records and calculated a phenotype risk score (PheScore) to indicate patient-phenotype similarity. Patients meeting predefined PheScore criteria but lacking diagnosis codes were identified. Clinically knowledgeable experts adjudicated randomly selected patients per phenotype as Positive, Possibly Positive, or Negative.

RESULTS

Our approach indicated that 5.3% of patients lacked a diagnosis for T2DM, 4.5% for dementia, 2.2% for prostate cancer, and 0.2% for sensorineural hearing loss. The expert review indicated 100% precision (for Possibly Positive or Positive cases) for dementia and sensorineural hearing loss, and 90.0% and 85.0% precision for T2DM and prostate cancer, respectively. Excluding Possibly Positive cases, the precision for T2DM and prostate cancer was 88.9% and 81.3%, respectively.

CONCLUSIONS

Leveraging clinical terminologies incorporated by PheMAP can effectively identify patients with phenotypes who lack EHR diagnosis codes, thereby enhancing phenotyping quality and related research reliability.

Identification of Transdiagnostic Childhood Externalizing Pathology Within an Electronic Medical Records Database and Application to the Analysis of Rare Copy Number Variation

Externalizing traits and behaviors are broadly defined by impairments in self-regulation and impulse control that typically begin in childhood and adolescence. Externalizing behaviors, traits, and symptoms span a range of traditional psychiatric diagnostic categories. In this study, we sought to generate an algorithm that could reliably identify transdiagnostic childhood-onset externalizing cases and controls within a university hospital electronic health record (EHR) database. Within the Vanderbilt University Medical Center (VUMC) EHR, our algorithm identified cases with a clinician-validated positive predictive value of 90% and controls with a negative predictive value of 88%. In individuals of genetically defined European ancestry (CEU-clustered; Ncase = 487, Ncontrol = 5638), case status was significantly associated with psychiatric comorbidity and with elevated externalizing polygenic scores (OR: 1.20; 95% CI: 1.09-1.33; p = 1.14 × 10-3; based on published genome-wide association data). To test whether our cohort definitions could be applied to generate novel genetic insights, we examined rare (allele frequency < 0.5%) copy number variation. An association (OR: 9.70; CI: 3.24-29.0) was identified in the CEU-clustered cohort on chromosome 2 (chr2: 45,408,678-45,551,530; duplication), although the statistical strength of this association was modest (p = 0.052). We also examined the role of an externalizing burden score based on the number of externalizing diagnoses present in cases and found similar results to our case-control analysis. This analysis identified several other statistically significant CNV region associations. This study provides a framework for identifying childhood externalizing case-control cohorts within an EHR. Future work should validate this framework within other health systems. A broadly applicable algorithm, like this one, may allow for detection of rare outcomes or outcomes in populations historically excluded from genomic research through meta-analysis of data across health care systems.

Clinical Characteristics associated with functional seizures in individuals with psychosis

BACKGROUND AND HYPOTHESIS

Functional seizures (FS) are episodes characterized by seizure-like events that are not caused by hypersynchronous neuronal activity. Prior studies have suggested an increased prevalence of psychotic disorders among patients with FS, but results have been inconsistent. We hypothesize that FS are associated with psychosis and that among patients with psychosis, the presence of FS may influence patient clinical characteristics, mortality, and medical resource utilization.

STUDY DESIGN

The association between FS and psychosis was assessed using electronic health records data from a total of 761,848 individuals receiving care at Vanderbilt University Medical Center between 1989 and 2023. Analyses of the association between FS and psychiatric outcomes, sexual trauma, healthcare utilization, and other clinical comorbidities were conducted in a subset of 5219 patients with psychosis.

STUDY RESULTS

Odds of FS were elevated among patients with psychosis compared to controls (OR = 10.09, 95 % CI = 8.40-12.13). Among patients with psychosis, those with FS exhibited higher rates of suicidality (OR = 2.18 95 % CI = 1.50-3.17), catatonia (OR = 2.15, 95 % CI = 1.33-3.45), sexual trauma history (OR = 2.93, 95 % CI = 2.00-4.29) and had a greater number of antipsychotic trials (4.63 versus 3.37, beta = 1.23, SE = 0.18, adjusted p < 0.001) than those without FS. Furthermore, patients with comorbid FS had more hospital presentations at one, three, five, and ten years after receiving a psychosis diagnosis (adjusted p < 0.001).

CONCLUSIONS

FS are more common among patients with psychosis and are associated with increased healthcare utilization as well as an increased prevalence of suicidality, catatonia, and certain psychiatric and medical comorbidities.

A phenome-wide association study of uterine fibroids reveals a marked burden of comorbidities

BACKGROUND

The burden of comorbidities in those with uterine fibroids compared to those without fibroids is understudied. We performed a phenome-wide association study to systematically assess the association between fibroids and other conditions.

METHODS

Vanderbilt University Medical Center's Synthetic Derivative and Geisinger Health System Database, two electronic health record databases, were used for discovery and validation. Non-Hispanic Black and White females were included. Fibroid cases were identified through a previously validated algorithm. Race-stratified and multi-population phenome-wide association analyses, adjusting for age and body mass index, were performed before statistically significant, validated results were meta-analyzed.

RESULTS

There were 52,295 and 26,918 (9022 and 10,232 fibroid cases) females included in discovery and validation analyses. In multi-population meta-analysis, 389 conditions were associated with fibroid risk, with evidence of enrichment of circulatory, dermatologic, genitourinary, musculoskeletal, and sense organ conditions. The strongest associations within and across racial groups included conditions previously associated with fibroids. Numerous novel diagnoses, including cancers in female genital organs, were tied to fibroid status.

CONCLUSIONS

Overall, individuals with fibroids have a marked increase in comorbidities compared to those without fibroids. This approach to evaluate the health context of fibroids highlights the potential to understand fibroid etiology through studying the common biology of comorbid diagnoses and through disease networks.

A quantitative analysis of the use of anonymization in biomedical research

Anonymized biomedical data sharing faces several challenges. This systematic review analyzes 1084 PubMed-indexed studies (2018-2022) using anonymized biomedical data to quantify usage trends across geographic, regulatory, and cultural regions to identify effective approaches and inform implementation agendas. We identified a significant yearly increase in such studies with a slope of 2.16 articles per 100,000 when normalized against the total number of PubMed-indexed articles (p = 0.021). Most studies used data from the US, UK, and Australia (78.2%). This trend remained when normalized by country-specific research output. Cross-border sharing was rare (10.5% of studies). We identified twelve common data sources, primarily in the US (seven) and UK (three), including commercial (seven) and public entities (five). The prevalence of anonymization in the US, UK, and Australia suggests their practices could guide broader adoption. Rare cross-border anonymized data sharing and differences between countries with comparable regulations underscore the need for global standards.

Rare genetic variation in PTPRB is associated with central serous chorioretinopathy, varicose veins and glaucoma

Central serous chorioretinopathy is an eye disease characterized by fluid buildup under the central retina whose etiology is not well understood. Abnormal choroidal veins in central serous chorioretinopathy patients have been shown to have similarities with varicose veins. To identify potential mechanisms, we analyzed genotype data from 1,477 patients and 455,449 controls in FinnGen. We identified an association for a low-frequency (allele frequency = 0.5%) missense variant (rs113791087) in PTPRB, the gene encoding vascular endothelial protein tyrosine phosphatase (odds ratio=2.85, P = 4.5 × 10-9). This was confirmed in a meta-analysis of 2,452 patients and 865,767 controls from 4 studies (odds ratio=3.06, P = 7.4 × 10-15). Rs113791087 was associated with a 56% higher prevalence of retinal abnormalities (35.3% vs 22.6%, P = 8.0 × 10-4) in 708 UK Biobank participants and, surprisingly, with increased risk of varicose veins (odds ratio=1.31, P = 2.3 × 10-11) and reduced risk of glaucoma (odds ratio=0.82, P = 6.9 × 10-9). Predicted loss-of-function variants in PTPRB, though rare in number, were associated with central serous chorioretinopathy in All of Us (odds ratio=17.09, P = 0.018). These findings highlight the significance of vascular endothelial protein tyrosine phosphatase in diverse ocular and systemic veno-vascular diseases.

Expanding scope of genetic studies in the era of biobanks

Biobanks have become pivotal in genetic research, particularly through genome-wide association studies (GWAS), driving transformative insights into the genetic basis of complex diseases and traits through the integration of genetic data with phenotypic, environmental, family history, and behavioral information. This review explores the distinct design and utility of different biobanks, highlighting their unique contributions to genetic research. We further discuss the utility and methodological advances in combining data from disease-specific study or consortia with that of biobanks, especially focusing on summary statistics based meta-analysis. Subsequently we review the spectrum of additional advantages offered by biobanks in genetic studies in representing population differences, calibration of polygenic scores, assessment of pleiotropy and improving post-GWAS in silico analyses. Advances in sequencing technologies, particularly whole-exome and whole-genome sequencing, have further enabled the discovery of rare variants at biobank scale. Among recent developments, the integration of large-scale multi-omics data especially proteomics and metabolomics, within biobanks provides deeper insights into disease mechanisms and regulatory pathways. Despite challenges like ascertainment strategies and phenotypic misclassification, biobanks continue to evolve, driving methodological innovation and enabling precision medicine. We highlight the contributions of biobanks to genetic research, their growing integration with multi-omics, and finally discuss their future potential for advancing healthcare and therapeutic development.

Absence of nonfatal suicidal behavior preceding suicide death reveals differences in clinical risks

Nonfatal suicidal behavior is the most robust predictor of suicide death. However, only ∼10 % of those who survive an attempt go on to die by suicide. Moreover, ∼50 % of suicide deaths occur in the absence of prior known attempts, suggesting risks other than nonfatal suicide attempt need to be identified to help prevent suicide mortality. We studied data from 4,000 population-ascertained suicide deaths and 26,191 population controls to improve understanding of suicide deaths without prior nonfatal attempts. This study included 2,253 suicide deaths and 3,375 controls with evidence of nonfatal suicidal ideation or behaviors (SUI_SI/SB and CTL_SI/SB) from diagnostic codes and natural language processing of electronic health records notes. Characteristics of these groups were compared to 1,669 suicides with no prior nonfatal SI/SB (SUI_None) and 22,816 controls with no lifetime suicidality (CTL_None). The SUI_None and CTL_None groups had fewer overall diagnoses and were older than SUI_SI/SB and CTL_SI/SB. Mental health diagnoses were far less common in both the SUI_None and CTL_None groups; mental health problems were far less associated with suicide death than with presence of SI/SB. Physical health diagnoses were conversely more often associated with risk of suicide death than with presence of SI/SB. Pending replication, results indicate highly significant clinical differences among suicide deaths with versus without prior nonfatal SI/SB, and suggest that, for a substantial number of individuals at risk for suicide mortality, history of SI/SB does not serve as an effective clinical marker of risk.

Sex-specific Mendelian randomization phenome-wide association study of basal metabolic rate

Observationally, higher basal metabolic rate (BMR) is associated with metabolism-related disorders, cancer, aging, and mortality. In this Mendelian randomization (MR) phenome-wide association study, using two-sample MR methods, we systematically and comprehensively investigated the health effects of genetically predicted BMR across the phenome sex-specifically. We obtained sex-specific genetic variants strongly (p < 5 × 10- 8) and independently (r2 < 0.001) predicting BMR from the UK Biobank and applied them to over 1,000 phenotypes within the same study. We combined genetic variant-specific Wald estimates using inverse-variance weighting, supplemented by sensitivity analysis. We used a false-discovery rate correction to allow for multiple comparisons as well as multivariable MR adjusted for body mass index and testosterone to investigate the independent effects of BMR on phenotypes with significant univariable associations. We obtained 217/219 genetic variants predicting BMR and applied them to 1,150/1,242 phenotypes in men/women, respectively. BMR was associated with 190/270 phenotypes in univariable analysis and 122/123 phenotypes in multivariable analysis in men/women. Examples of robust associations in multivariable analysis included those with neoplasms, diseases of the circulatory system, and growth and reproductive investment. In conclusion, BMR might affect a wide range of health-related outcomes. The underlying mechanisms and interactions between phenotypes warrant further study, as BMR is modifiable.

Phenome-wide comorbidity network analysis reveals clinical risk patterns in enthesopathy and enthesitis

Background

Enthesopathy and enthesitis, including rotator cuff disease and other tendon disorders, represent a heterogeneous group of musculoskeletal conditions with complex etiologies. Understanding how systemic health profiles influence their onset remains a critical challenge in musculoskeletal medicine.

Methods

We conducted a large-scale, phenome-wide comorbidity analysis using longitudinal electronic health records (EHR) from 432,757 UK Biobank participants. Incident cases of peripheral enthesopathies were compared to controls across 434 baseline disease phenotypes. A directed ego network was constructed to link significantly associated comorbidities to the target condition using odds ratio-based associations. Unsupervised clustering via UMAP and DBSCAN identified data-driven comorbidity clusters, which were consolidated into unified endotypes-interpreted as distinct systemic profiles contributing to disease risk. Additionally, metapath-based trajectory analysis was applied to uncover temporally structured multimorbidity chains leading to disease onset.

Results

We identified 183 baseline conditions significantly associated with the future development of enthesopathy (FDR < 0.05). Network clustering revealed eight comorbidity clusters, which were consolidated into four unified endotypes: Metabolic-Psychosomatic, Inflammatory-Multisystem, Mechanical-Injury-driven, and Aging-Intervention-related. Metapath analysis uncovered common three-step disease trajectories, such as metabolic-infectious-musculoskeletal and inflammatory skin-to-joint progressions, highlighting potential mechanistic pathways. These endotypes showed diverse clinical features but shared biological coherence, suggesting that different systemic health profiles can converge to drive tendon-related disease.

Conclusions

This study introduces a scalable framework for identifying systemic multimorbidity patterns underlying enthesopathy and enthesitis using phenome-wide comorbidity networks. By integrating network clustering and metapath analysis, we uncover interpretable, data-driven endotypes that may inform individualized risk assessment and targeted care strategies. These findings contribute to the growing field of biobank-scale disease modeling and offer a foundation for precision approaches in musculoskeletal medicine.

GeneSetPheno: a web application for the integration, summary, and visualization of gene and variant-phenotype associations across gene sets

Motivation

The comprehensive study of genotype-phenotype relationships requires the integration of multiple data types to "triangulate" signals and derive meaningful biological conclusions. Large-scale biobanks and public resources generate a wealth of comprehensive results, facilitating the discovery of associations between genes or genetic variants and multiple phenotypes. However, analyzing these data across resources presents several challenges, including limited flexibility in gene set analysis, the integration of multipe databases, and the need for effective data visualization to aid interpretation.

Results

GeneSetPheno is a user-friendly graphical interface that integrates, summarizes, and visualizes gene and variant-phenotype associations across genomic resources. It allows users to explore interrelationships between genetic variants and phenotypes, offering insights into the genetic factors driving phenotypic variation within user-defined gene sets. GeneSetPheno also supports comparisons across gene sets to identify shared or unique genetic variants, phenotypic associations, biological pathways, and potential gene-gene interactions. GeneSetPheno is a free and highly configurable tool for exploring the complex relationships between gene sets, genetic variants, and phenotypes. Target users include molecular biologists and clinicians who wish to explore a gene or gene set of particular interest.

Availability and implementation

GeneSetPheno is freely accessible at: https://shiny.wehi.edu.au/han.ji/GeneSetPheno/. The source code is available on GitHub at: https://github.com/bahlolab/GeneSetPheno.

Identification of patients at risk for pancreatic cancer in a 3-year timeframe based on machine learning algorithms

Early detection of pancreatic cancer (PC) remains challenging largely due to the low population incidence and few known risk factors. However, screening in at-risk populations and detection of early cancer has the potential to significantly alter survival. In this study, we aim to develop a predictive model to identify patients at risk for developing new-onset PC at two and a half to three year time frame. We used the Electronic Health Records (EHR) of a large medical system from 2000 to 2021 (N = 537,410). The EHR data analyzed in this work consists of patients' demographic information, diagnosis records, and lab values, which are used to identify patients who were diagnosed with pancreatic cancer and the risk factors used in the machine learning algorithm for prediction. We identified 73 risk factors of pancreatic cancer with the Phenome-wide Association Study (PheWAS) on a matched case-control cohort. Based on them, we built a large-scale machine learning algorithm based on EHR. A temporally stratified validation based on patients not included in any stage of the training of the model was performed. This model showed an AUROC at 0.742 [0.727, 0.757] which was similar in both the general population and in a subset of the population who has had prior cross-sectional imaging. The rate of diagnosis of pancreatic cancer in those in the top 1 percentile of the risk score was 6 folds higher than the general population. Our model leverages data extracted from a 6-month window of time in the electronic health record to identify patients at nearly sixfold higher than baseline risk of developing pancreatic cancer 2.5-3 years from evaluation. This approach offers an opportunity to define an enriched population entirely based on static data, where current screening may be recommended.

Pierre CL, Missfeldt Sanches T, Bimschleger HV, Gao J, Cheng R, Nguyen K, Holl KL, Polesskaya O, Ishiwari K, Chen H, Robinson TE, Flagel SB, Solberg Woods LC, Palmer AA, Meyer PJ. Genetic Loci Influencing Cue-Reactivity in Heterogeneous Stock Rats

Addiction vulnerability is associated with the tendency to attribute incentive salience to reward predictive cues. Both addiction and the attribution of incentive salience are influenced by environmental and genetic factors. To characterize the genetic contributions to incentive salience attribution, we performed a genome-wide association study (GWAS) in a cohort of 1596 heterogeneous stock (HS) rats. Rats underwent a Pavlovian conditioned approach task that characterized the responses to food-associated stimuli ("cues"). Responses ranged from cue-directed "sign-tracking" behavior to food-cup directed "goal-tracking" behavior (12 measures, SNP heritability: 0.051-0.215). Next, rats performed novel operant responses for unrewarded presentations of the cue using the conditioned reinforcement procedure. GWAS identified 14 quantitative trait loci (QTLs) for 11 of the 12 traits across both tasks. Interval sizes of these QTLs varied widely. Seven traits shared a QTL on chromosome 1 that contained a few genes (e.g., Tenm4, Mir708) that have been associated with substance use disorders and other psychiatric disorders in humans. Other candidate genes (e.g., Wnt11, Pak1) in this region had coding variants and expression-QTLs in mesocorticolimbic regions of the brain. We also conducted a Phenome-Wide Association Study (PheWAS) on addiction-related behaviors in HS rats and found that the QTL on chromosome 1 was also associated with nicotine self-administration in a separate cohort of HS rats. These results provide a starting point for the molecular genetic dissection of incentive motivational processes and provide further support for a relationship between the attribution of incentive salience and drug abuse-related traits.

Linking epidemiology and genomics of maternal smoking during pregnancy in utero and in ageing: a population-based study using human foetuses and the UK Biobank cohort

BACKGROUND

Maternal smoking and foetal exposure to nicotine and other harmful chemicals in utero remains a serious public health issue with little knowledge about the underlying genetics and consequences of maternal smoking in ageing individuals. Here, we investigated the epidemiology and genomic architecture of maternal smoking in a middle-aged population and compare the results to effects observed in the developing foetus.

METHODS

In the current project, we included 351,562 participants from the UK Biobank (UKB) and estimated exposure to maternal smoking status during pregnancy through self-reporting from the UKB participants about the mother's smoking status around their birth. In addition, we analysed 64 foetal liver transcriptomic expression datasets collected from women seeking elective pregnancy terminations. Foetal maternal smoking exposure was confirmed through measurement of foetal plasma cotinine levels.

FINDINGS

Foetal exposure to maternal smoking had a greater impact on males than females, with more differentially expressed genes in liver tissue (3313 vs. 1163) and higher liver pathway activation. In the UKB, maternal smoking exposure was linked to an unhealthy lifestyle, lower education, and liver damage. In a genome-wide analysis in the UKB, we leveraged the shared genetic basis between affected offspring and their mothers and identified five genome-wide significant regions. We found a low heritability of the trait (∼4%) and implicated several disease-related genes in a transcriptome-wide association study. Maternal smoking increased all-cause mortality risk (Hazard ratio and 95% CI: 1.10 [1.04; 1.16], P = 4.04 × 10-4), which was attenuated in non-smoking males.

INTERPRETATION

Although male foetuses are more affected than females by maternal smoking in pregnancy, this effect was largely reduced in middle-aged individuals. Importantly, our results highlight that the overall 10% increased mortality due to maternal smoking in pregnancy was greatly attenuated in non-smokers. This study demonstrates the importance of campaigns promoting offspring smoking prevention in families where the parent(s) smoke.

FUNDING

Funding for this project was provided by the University of Aberdeen, the Science Initiative Panel of the Institute of Medical Science, the UK Medical Research Council, the Seventh Framework Programme of the European Union under Grant Agreement 212885 (REEF), NHS Grampian Endowments grants and the European Commission Horizon Europe research grant Agreement 101094099 (INITIALISE).

Longitudinal Analysis of Electronic Health Records Reveals Medical Conditions Associated with Subsequent Alzheimer's Disease Development

Background

Several health conditions are known to increase the risk of Alzheimer's disease (AD). We aim to systematically identify medical conditions that are associated with subsequent development of AD by leveraging the growing resources of electronic health records (EHRs).

Methods

This retrospective cohort study used de-identified EHRs from two independent databases (MarketScan and VUMC) with 153 million individuals to identify AD cases and age- and gender-matched controls. By tracking their EHRs over a 10-year window before AD diagnosis and comparing the EHRs between AD cases and controls, we identified medical conditions that occur more likely in those who later develop AD. We further assessed the genetic underpinnings of these conditions in relation to AD genetics using data from two large-scale biobanks (BioVU and UK Biobank, total N=450,000).

Results

We identified 43,508 AD cases and 419,455 matched controls in MarketScan, and 1,320 AD cases and 12,720 matched controls in VUMC. We detected 406 and 102 medical phenotypes that are significantly enriched among the future AD cases in MarketScan and VUMC databases, respectively. In both EHR databases, mental disorders and neurological disorders emerged as the top two most enriched clinical categories. More than 70 medical phenotypes are replicated in both EHR databases, which are dominated by mental disorders (e.g., depression), neurological disorders (e.g., sleep orders), circulatory system disorders (e.g. cerebral atherosclerosis) and endocrine/metabolic disorders (e.g., type 2 diabetes). We identified 19 phenotypes that are either associated with individual risk variants of AD or a polygenic risk score of AD.

Conclusions

In this study, analysis of longitudinal EHRs from independent large-scale databases enables robust identification of health conditions associated with subsequent development of AD, highlighting potential opportunities of therapeutics and interventions to reduce AD risk.

An in-depth review of AI-powered advancements in cancer drug discovery

The convergence of artificial intelligence (AI) and genomics is redefining cancer drug discovery by facilitating the development of personalized and effective therapies. This review examines the transformative role of AI technologies, including deep learning and advanced data analytics, in accelerating key stages of the drug discovery process: target identification, drug design, clinical trial optimization, and drug response prediction. Cutting-edge tools such as DrugnomeAI and PandaOmics have made substantial contributions to therapeutic target identification, while AI's predictive capabilities are driving personalized treatment strategies. Additionally, advancements like AlphaFold highlight AI's capacity to address intricate challenges in drug development. However, the field faces significant challenges, including the management of large-scale genomic datasets and ethical concerns surrounding AI deployment in healthcare. This review underscores the promise of data-centric AI approaches and emphasizes the necessity of continued innovation and interdisciplinary collaboration. Together, AI and genomics are charting a path toward more precise, efficient, and transformative cancer therapeutics.

Towards Scalable Biomarker Discovery in Posttraumatic Stress Disorder: Triangulating Genomic and Phenotypic Evidence from a Health System Biobank

Importance

Biomarkers can potentially improve the diagnosis, monitoring, and treatment of posttraumatic stress disorder (PTSD). However, PTSD biomarkers that are scalable and easily integrated into real-world clinical settings have not been identified.

Objective

To triangulate phenotypic and genomic evidence from a health system biobank with a goal of identifying scalable and clinically relevant biomarkers for PTSD.

Design setting and participants

The analysis was conducted between June to November 2024 using genomic samples and laboratory test results recorded in the Mass General Brigham (MGB) Health System. The analysis included 23,743 European ancestry participants from the nested MGB Biobank study.

Exposures

The first exposure was polygenic risk score (PRS) for PTSD, calculated using the largest available European ancestry genome-wide association study (GWAS), employing a Bayesian polygenic scoring method. The second exposure was a clinical diagnosis of PTSD, determined by the presence of two or more qualifying PTSD phecodes in the longitudinal electronic health records (EHR).

Main outcomes and measures

The primary outcomes were the inverse normal quantile transformed, median lab values of 241 laboratory traits with non-zero h 2 SNP estimates.

Results

Sixteen unique laboratory traits across the cardiometabolic, hematologic, hepatic, and immune systems were implicated in both genomic and phenotypic lab-wide association scans (LabWAS). Two-sample Mendelian randomization analyses provided evidence of potential unidirectional causal effects of PTSD liability on five laboratory traits.

Conclusion and relevance

These findings demonstrate the potential of a triangulation approach to uncover scalable and clinically relevant biomarkers for PTSD.

Outcome adaptive propensity score methods for handling censoring and high-dimensionality: Application to insurance claims

Propensity scores are commonly used to reduce the confounding bias in non-randomized observational studies for estimating the average treatment effect. An important assumption underlying this approach is that all confounders that are associated with both the treatment and the outcome of interest are measured and included in the propensity score model. In the absence of strong prior knowledge about potential confounders, researchers may agnostically want to adjust for a high-dimensional set of pre-treatment variables. As such, variable selection procedure is needed for propensity score estimation. In addition, studies show that including variables related to treatment only in the propensity score model may inflate the variance of the treatment effect estimators, while including variables that are predictive of only the outcome can improve efficiency. In this article, we propose to incorporate outcome-covariate relationship in the propensity score model by including the predicted binary outcome probability as a covariate. Our approach can be easily adapted to an ensemble of variable selection methods, including regularization methods and modern machine-learning tools based on classification and regression trees. We evaluate our method to estimate the treatment effects on a binary outcome, which is possibly censored, across multiple treatment groups. Simulation studies indicate that incorporating outcome probability for estimating the propensity scores can improve statistical efficiency and protect against model misspecification. The proposed methods are applied to a cohort of advanced-stage prostate cancer patients identified from a private insurance claims database for comparing the adverse effects of four commonly used drugs for treating castration-resistant prostate cancer.

Distinguishing clinical and genetic risk factors for suicidal ideation and behavior in a diverse hospital population

Suicidal ideation (SI) and behavior (SB) are major public health concerns, but risk factors for their development and progression are poorly understood. We used ICD codes and a natural language processing algorithm to identify individuals in a hospital biobank with SI-only, SB, and controls without either. We compared the profiles of SB and SI-only patients to controls, and each other, using phenome-wide association studies (PheWAS) and polygenic risk scores (PRS). PheWAS identified many risk factors for SB and SI-only, plus specific psychiatric disorders which may be involved in progression from SI-only to SB. PRS for suicide attempt were only associated with SB, and even after accounting for psychiatric disorder PRS. SI PRS were only associated with SI-only, although not after accounting for psychiatric disorder PRS. These findings advance understanding of distinct genetic and clinical risk factors for SB and SI-only, which will aid in early detection and intervention efforts.

Interstitial Cystitis: a phenotype and rare variant exome sequencing study: Interstitial Cystitis: a phenotype and exome sequencing study

Interstitial cystitis/bladder pain syndrome (IC/BPS) is a poorly understood and underdiagnosed syndrome of chronic bladder/pelvic pain with urinary frequency and urgency. Though IC/BPS can be hereditary, little is known of its genetic etiology. Using the eMERGE data, we confirmed known phenotypic associations such as gastroesophageal reflux disease and irritable bowel syndrome and detected new associations, including osteoarthrosis/osteoarthritis and Barrett's esophagus. An exome wide ultra-rare variants analysis in 348 IC/BPS and 11,981 controls extended the previously reported association with ATP2C1 and ATP2A2, implicated in Mendelian desquamating skin disorders, but did not provide evidence for other previously proposed pathogenic pathways such as bladder development, nociception or inflammation. Pathway analysis detected new associations with "anaphase-promoting complex-dependent catabolic process", the "regulation of MAPK cascade" and "integrin binding". These findings suggest perturbations in biological networks for epithelial integrity and cell cycle progression in IC/BPS pathogenesis, and provide a roadmap for its future investigation.

Gene and phenome-based analysis of the shared genetic architecture of eye diseases

While many eye disorders are linked through defects in vascularization and optic nerve degeneration, genetic correlation studies have yielded variable results despite shared features. For example, glaucoma and myopia both share optic neuropathy as a feature, but genetic correlation studies demonstrated minimal overlap. By leveraging electronic health record (EHR) resources that contain genetic variables such as genetically predicted gene expression (GPGE), researchers have the potential to improve the identification of shared genetic drivers of disease by incorporating knowledge of shared features to identify disease-causing mechanisms. In this study, we examined shared genetic architecture across eye diseases. Our gene-based approach used transcriptome-wide association methods to identify shared transcriptomic profiles across eye diseases within BioVU, Vanderbilt University Medical Center's (VUMC's) EHR-linked biobank. Our phenome-based approach leveraged phenome-wide association studies (PheWASs) to identify eye disease comorbidities. Using the beta estimates from the significantly associated comorbidities, we constructed a phenotypic risk score (PheRS) representing a weighted sum of an individual's eye disease comorbidities. This PheRS is predictive of eye disease status and associated with the altered GPGE of significant genes in an independent population. The implementation of both gene- and phenome-based approaches can expand genetic associations and shed greater insight into the underlying mechanisms of shared genetic architecture across eye diseases.

Exploring beyond diagnoses in electronic health records to improve discovery: a review of the phenome-wide association study

Objective

The phenome-wide association study (PheWAS) systematically examines the phenotypic spectrum extracted from electronic health records (EHRs) to uncover correlations between phenotypes and exposures. This review explores methodologies, highlights challenges, and outlines future directions for EHR-driven PheWAS.

Materials and Methods

We searched the PubMed database for articles spanning from 2010 to 2023, and we collected data regarding exposures, phenotypes, cohorts, terminologies, replication, and ancestry.

Results

Our search yielded 690 articles. Following exclusion criteria, we identified 291 articles published between January 1, 2010, and December 31, 2023. A total number of 162 (55.6%) articles defined phenomes using phecodes, indicating that research is reliant on the organization of billing codes. Moreover, 72.8% of articles utilized exposures consisting of genetic data, and the majority (69.4%) of PheWAS lacked replication analyses.

Discussion

Existing literature underscores the need for deeper phenotyping, variability in PheWAS exposure variables, and absence of replication in PheWAS. Current applications of PheWAS mainly focus on cardiovascular, metabolic, and endocrine phenotypes; thus, applications of PheWAS in uncommon diseases, which may lack structured data, remain largely understudied.

Conclusions

With modern EHRs, future PheWAS should extend beyond diagnosis codes and consider additional data like clinical notes or medications to create comprehensive phenotype profiles that consider severity, temporality, risk, and ancestry. Furthermore, data interoperability initiatives may help mitigate the paucity of PheWAS replication analyses. With the growing availability of data in EHR, PheWAS will remain a powerful tool in precision medicine.

The interplay of sex and genotype in disease associations: a comprehensive network analysis in the UK Biobank

BACKGROUND

Disease comorbidities and longer-term complications, arising from biologically related associations across phenotypes, can lead to increased risk of severe health outcomes. Given that many diseases exhibit sex-specific differences in their genetics, our objective was to determine whether genotype-by-sex (GxS) interactions similarly influence cross-phenotype associations. Through comparison of sex-stratified disease-disease networks (DDNs)-where nodes represent diseases and edges represent their relationships-we investigate sex differences in patterns of polygenicity and pleiotropy between diseases.

RESULTS

Using UK Biobank summary statistics, we built male- and female-specific DDNs for 103 diseases. This revealed that male and female diseasomes have similar topology and central diseases (e.g., hypertensive, chronic respiratory, and thyroid-based disorders), yet some phenotypes exhibit sex-specific influence in cross-phenotype associations. Multiple sclerosis and osteoarthritis are central only in the female DDN, while cardiometabolic diseases and skin cancer are more prominent in the male DDN. Edge comparison indicated similar shared genetics between the two graphs relative to a random model of disease association, though notable discrepancies in embedding distances and clustering patterns imply a more expansive genetic influence on multimorbidity risk for females than males. Analysis of pleiotropic contributions of two sexually-dimorphic single-nucleotide polymorphisms related to thyroid disorders further validated a distinct genetic architecture across sexes that influences associations, confirmed through examination of corresponding gene expression profiles from the GTEx Portal.

CONCLUSIONS

Our analysis affirms the presence of GxS interactions in cross-phenotype associations, emphasizing the need to investigate the role of sex in disease onset and its importance in biomedical discovery and precision medicine research.

Unsupervised deep learning of electrocardiograms enables scalable human disease profiling

The 12-lead electrocardiogram (ECG) is inexpensive and widely available. Whether conditions across the human disease landscape can be detected using the ECG is unclear. We developed a deep learning denoising autoencoder and systematically evaluated associations between ECG encodings and ~1,600 Phecode-based diseases in three datasets separate from model development, and meta-analyzed the results. The latent space ECG model identified associations with 645 prevalent and 606 incident Phecodes. Associations were most enriched in the circulatory (n = 140, 82% of category-specific Phecodes), respiratory (n = 53, 62%) and endocrine/metabolic (n = 73, 45%) categories, with additional associations across the phenome. The strongest ECG association was with hypertension (p < 2.2×10-308). The ECG latent space model demonstrated more associations than models using standard ECG intervals, and offered favorable discrimination of prevalent disease compared to models comprising age, sex, and race. We further demonstrate how latent space models can be used to generate disease-specific ECG waveforms and facilitate individual disease profiling.

A phenome-wide association study of polygenic scores for selected childhood cancer: Results from the UK Biobank

The aim of this study was to scan phenotypes in adulthood associated with polygenic risk scores (PRS) for childhood cancers with well-articulated genetic architectures-acute lymphoblastic leukemia (ALL), Ewing sarcoma, and neuroblastoma-to examine genetic pleiotropy. Furthermore, we aimed to determine which SNPs could drive associations. Per-SNP summary statistics were extracted for PRS calculation. Participants with white British ancestry were exclusively included for analyses. SNPs were queried from the UK Biobank genotype imputation data. Records from the cancer registry, death registry, and inpatient diagnoses were abstracted for phenome-wide scans. Firth logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) alongside corresponding p values, adjusting for age at recruitment and sex. A total of 244,332 unrelated white British participants were included. We observed a significant association between ALL-PRS and ALL (OR: 1.20e+24, 95% CI: 9.08e+14-1.60e+33). In addition, we observed a significant association between high-risk neuroblastoma PRS and nonrheumatic aortic valve disorders (OR: 43.9, 95% CI: 7.42-260). There were no significant phenotype associations with Ewing sarcoma and neuroblastoma PRS. Regarding individual SNPs, rs17607816 increased the risk of ALL (OR: 6.40, 95% CI: 3.26-12.57). For high-risk neuroblastoma, rs80059929 elevated the risk of atrioventricular block (OR: 3.04, 95% CI: 1.85-4.99). Our findings suggest that individuals with genetic susceptibility to ALL may face a lifelong risk for developing ALL, along with a genetic pleiotropic association between high-risk neuroblastoma and circulatory diseases.

PWAS Hub: exploring gene-based associations of complex diseases with sex dependency

The Proteome-Wide Association Study (PWAS) is a protein-based genetic association approach designed to complement traditional variant-based methods like GWAS. PWAS operates in two stages: first, machine learning models predict the impact of genetic variants on protein-coding genes, generating effect scores. These scores are then aggregated into a gene-damaging score for each individual. This score is then used in case-control statistical tests to significantly link to specific phenotypes. PWAS Hub (v1.2) is a user-friendly platform that facilitates the exploration of gene-disease associations using clinical and genetic data from the UK Biobank (UKB), encompassing 500k individuals. PWAS Hub reports on 819 diseases and phenotypes determined by PheCode and ICD-10 clinical codes, each with a minimum of 400 affected individuals. PWAS-derived gene associations were reported for 72% of the tested phenotypes. The PWAS Hub also analyzes gene associations separately for males and females, considering sex-specific genetic effects, inheritance patterns (dominant and recessive), and gene pleiotropy. We illustrated the utility of the PWAS Hub for primary (essential) hypertension (I10), type 2 diabetes mellitus (E11), and specified haematuria (R31) that showed sex-dependent genetic signals. The PWAS Hub, available at pwas.huji.ac.il, is a valuable resource for studying genetic contributions to common diseases and sex-specific effects.

GWAS highlights the neuronal contribution to multiple sclerosis susceptibility

Multiple Sclerosis (MS) is a chronic inflammatory and neurodegenerative disease affecting the brain and spinal cord. Genetic studies have identified many risk loci, that were thought to primarily impact immune cells and microglia. Here, we performed a multi-ancestry genome-wide association study with 20,831 MS and 729,220 control participants, identifying 236 susceptibility variants outside the Major Histocompatibility Complex, including four novel loci. We derived a polygenic score for MS and, optimized for European ancestry, it is informative for African-American and Latino participants. Integrating single-cell data from blood and brain tissue, we identified 76 genes affected by MS risk variants. Notably, while T cells showed the strongest enrichment, inhibitory neurons emerged as a key cell type. The expression of IL7 and STAT3 are affected only in inhibitory neurons, highlighting the importance of neuronal and glial dysfunction in MS susceptibility.

Constructing a multi-ancestry polygenic risk score for uterine fibroids using publicly available data highlights need for inclusive genetic research

Uterine leiomyomata, or fibroids, are common gynecological tumors causing pelvic and menstrual symptoms that can negatively affect quality of life and child-bearing desires. As fibroids grow, symptoms can intensify and lead to invasive treatments that are less likely to preserve fertility. Identifying individuals at highest risk for fibroids can aid in access to earlier diagnoses. Polygenic risk scores (PRS) quantify genetic risk to identify those at highest risk for disease. Utilizing the PRS software PRS-CSx and publicly available genome-wide association study (GWAS) summary statistics from FinnGen and Biobank Japan, we constructed a multi-ancestry (META) PRS for fibroids. We validated the META PRS in two cross-ancestry cohorts. In the cross-ancestry Electronic Medical Record and Genomics (eMERGE) Network cohort, the META PRS was significantly associated with fibroid status and exhibited 1.11 greater odds for fibroids per standard deviation increase in PRS (95% confidence interval [CI]: 1.05 - 1.17, p = 5.21x10-5). The META PRS was validated in two BioVU cohorts: one using ICD9/ICD10 codes and one requiring imaging confirmation of fibroid status. In the ICD cohort, a standard deviation increase in the META PRS increased the odds of fibroids by 1.23 (95% CI: 1.15 - 1.32, p = 9.68x10-9), while in the imaging cohort, the odds increased by 1.26 (95% CI: 1.18 - 1.35, p = 2.40x10-11). We subsequently constructed single ancestry PRS for FinnGen (European ancestry [EUR]) and Biobank Japan (East Asian ancestry [EAS]) using PRS-CS and discovered a nominally significant association in the eMERGE cohort within fibroids and EAS PRS but not EUR PRS (95% CI: 1.09 - 1.20, p = 1.64x10-7). These findings highlight the strong predictive power of multi-ancestry PRS over single ancestry PRS. This study underscores the necessity of diverse population inclusion in genetic research to ensure precision medicine benefits all individuals equitably.

A bioinformatics toolbox to prioritize causal genetic variants in candidate regions

This review addresses the significant challenge of identifying causal genetic variants within quantitative trait loci (QTLs) for complex traits and diseases. Despite progress in detecting the ever-larger number of such loci, establishing causality remains daunting. We advocate for integrating bioinformatics and multiomics analyses to streamline the prioritization of candidate genes' variants. Our case study on the Pla2g4e gene, identified previously as a positional candidate obesity gene through genetic mapping and expression studies, demonstrates how applying multiomic data filtered through regulatory elements containing SNPs can refine the search for causative variants. This approach can yield results that guide more efficient experimental strategies, accelerating genetic research toward functional validation and therapeutic development.

Curation and validation of electronic medical record-based dementia diagnoses in the VA Million Veteran Program

BACKGROUND

The age distribution and diversity of the VA Million Veteran Program (MVP) cohort make it a valuable resource for studying the genetics of Alzheimer's disease (AD) and related dementias (ADRD).

OBJECTIVE

We present and evaluate the performance of several International Classification of Diseases (ICD) code-based classification algorithms for AD, ADRD, and dementia for use in MVP genetic studies and other studies using VA electronic medical record (EMR) data. These were benchmarked relative to existing ICD algorithms and AD-medication-identified cases.

METHODS

We used chart review of n = 103 MVP participants to evaluate diagnostic utility of the algorithms. Suitability for genetic studies was examined by assessing association with APOE ε4, the strongest genetic AD risk factor, in a large MVP cohort (n = 286 K).

RESULTS

The newly developed MVP-ADRD algorithm performed well, comparable to the existing PheCode dementia algorithm (Phe-Dementia) in terms of sensitivity (0.95 and 0.95) and specificity (0.65 and 0.70). The strongest APOE ε4 associations were observed in cases identified using MVP-ADRD and Phe-Dementia augmented with medication-identified cases (MVP-ADRD or medication, p = 3.6 ×10-290; Phe-Dementia or medication, p = 1.4 ×10-290). Performance was improved when cases were restricted to those with onset age ≥60.

CONCLUSIONS

We found that our MVP-developed ICD-based algorithms had good performance in chart review and generated strong genetic signals, especially after inclusion of medication-identified cases. Ultimately, our MVP-derived algorithms are likely to have good performance in the broader VA, and their performance may also be suitable for use in other large-scale EMR-based biobanks in the absence of definitive biomarkers such as amyloid-PET and cerebrospinal fluid biomarkers.

Exploring pleiotropy in Mendelian randomisation analyses: What are genetic variants associated with 'cigarette smoking initiation' really capturing?

Genetic variants used as instruments for exposures in Mendelian randomisation (MR) analyses may have horizontal pleiotropic effects (i.e., influence outcomes via pathways other than through the exposure), which can undermine the validity of results. We examined the extent of this using smoking behaviours as an example. We first ran a phenome-wide association study in UK Biobank, using a smoking initiation genetic instrument. From the most strongly associated phenotypes, we selected those we considered could either plausibly or not plausibly be caused by smoking. We examined associations between genetic instruments for smoking initiation, smoking heaviness and lifetime smoking and these phenotypes in UK Biobank and the Avon Longitudinal Study of Parents and Children (ALSPAC). We conducted negative control analyses among never smokers, including children. We found evidence that smoking-related genetic instruments were associated with phenotypes not plausibly caused by smoking in UK Biobank and (to a lesser extent) ALSPAC. We observed associations with phenotypes among never smokers. Our results demonstrate that smoking-related genetic risk scores are associated with unexpected phenotypes that are less plausibly downstream of smoking. This may reflect horizontal pleiotropy in these genetic risk scores, and we would encourage researchers to exercise caution this when using these and genetic risk scores for other complex behavioural exposures. We outline approaches that could be taken to consider this and overcome issues caused by potential horizontal pleiotropy, for example, in genetically informed causal inference analyses (e.g., MR) it is important to consider negative control outcomes and triangulation approaches, to avoid arriving at incorrect conclusions.

Sexual Trauma, Polygenic Scores, and Mental Health Diagnoses and Outcomes

Importance

Leveraging real-world clinical biobanks to investigate the associations between genetic and environmental risk factors for mental illness may help direct clinical screening efforts and evaluate the portability of polygenic scores across environmental contexts.

Objective

To examine the associations between sexual trauma, polygenic liability to mental health outcomes, and clinical diagnoses of schizophrenia, bipolar disorder, and major depressive disorder in a clinical biobank setting.

Design, Setting, and Participants

This genetic association study was conducted using clinical and genotyping data from 96 002 participants across hospital-linked biobanks located at Vanderbilt University Medical Center (VUMC), Nashville, Tennessee (including 58 262 individuals with high genetic similarity to the 1000 Genomes Project [1KG] Northern European from Utah reference population [1KG-EU-clustered] and 11 047 with high genetic similarity to the 1KG African-ancestry reference population of Yoruba in Ibadan, Nigeria [1KG-YRI-clustered]), and Mass General Brigham (MGB), Boston, Massachusetts (26 693 individuals with high genetic similarity to the combined European-ancestry superpopulation [1KG-EU-clustered]). Clinical data analyzed included diagnostic billing codes and clinical notes spanning from 1976 to 2023. Data analysis was performed from 2022 to 2024.

Exposures

Clinically documented sexual trauma disclosures and polygenic scores for schizophrenia, bipolar disorder, and major depressive disorder.

Main Outcomes and Measures

Diagnoses of schizophrenia, bipolar disorder, and major depressive disorder, determined by aggregating related diagnostic billing codes, were the dependent variables in logistic regression models including sexual trauma disclosure status, polygenic scores, and their interactions as the independent variables.

Results

Across the VUMC and MGB biobanks, 96 002 individuals were included in analyses (VUMC 1KG-EU-clustered: 33 011 [56.7%] female; median [range] age, 56.8 [10.0 to >89] years; MGB 1KG-EU-clustered: 14 647 [54.9%] female; median [range] age, 58.0 [10.0 to >89] years; VUMC 1KG-YRI-clustered: 6961 [63.0%] female; median [range] age, 44.6 [10.1 to >89] years). Sexual trauma history was associated with all mental health conditions across institutions (ORs ranged from 8.83 [95% CI, 5.50-14.18] for schizophrenia in the VUMC 1KG-YRI-clustered cohort to 17.65 [95% CI, 12.77-24.40] for schizophrenia in the VUMC 1KG-EU-clustered cohort). Sexual trauma history and polygenic scores jointly explained 3.8% to 8.8% of mental health phenotypic variance. Schizophrenia and bipolar disorder polygenic scores had greater associations with mental health outcomes in individuals with no documented disclosures of sexual trauma (schizophrenia interaction: OR, 0.70 [95% CI, 0.56-0.88]; bipolar disorder interaction: OR, 0.83 [95% CI, 0.74-0.94]).

Conclusions and Relevance

Sexual trauma and mental health polygenic scores, while correlated with one another, were independent and joint risk factors for severe mental illness in a large, diverse hospital biobank population. Furthermore, associations of schizophrenia and bipolar disorder polygenic scores with respective diagnoses were greater in those without disclosures, suggesting that genetic predisposition to mental illness as measured by polygenic scores may be less impactful in the presence of this severe environmental risk factor.

Phenome-wide association network demonstrates close connection with individual disease trajectories from the HUNT study

Disease networks offer a potential road map of connections between diseases. Several studies have created disease networks where diseases are connected either based on shared genes or Single Nucleotide Polymorphism (SNP) associations. However, it is still unclear to which degree SNP-based networks map to empirical, co-observed diseases within a different, general, adult study population spanning over a long time period. We created a SNP-based phenome-wide association network (PheNet) from a large population using the UK biobank phenome-wide association studies. Importantly, the SNP-associations are unbiased towards much studied diseases, adjusted for linkage disequilibrium, case/control imbalances, as well as relatedness. We map the PheNet to significantly co-occurring diseases in the Norwegian HUNT study population, and further, identify consecutively occurring diseases with significant ordering in occurrence, independent of age and gender in the PheNet. Our analysis reveals an overlap far larger than expected by chance between the two disease networks, with diseases typically connecting within their own category. Upon examining the sequential occurrence of diseases in the HUNT dataset, we find a giant component consisting of mostly cardiovascular disorders. This allows us to identify sequentially occurring diseases that are genetically linked and co-occur frequently, while also highlighting non-sequential diseases. Furthermore, we observe that survivors of severe cardiovascular diseases subsequently often face less severe conditions, but with a reduced time until their next fatal illness. The HUNT sub-PheNet showing both genetically and co-observed diseases offers an interesting framework to study groups of diseases and examine if they, in fact, are comorbidities. We find that the HUNT sub-PheNet offers the possibility to pinpoint exactly which mutation(s) constitute shared cause of the diseases. This could be of great benefit to both researchers and clinicians studying relationships between diseases.

PheWAS analysis on large-scale biobank data with PheTK

SUMMARY

With the rapid growth of genetic data linked to electronic health record (EHR) data in huge cohorts, large-scale phenome-wide association study (PheWAS) have become powerful discovery tools in biomedical research. PheWAS is an analysis method to study phenotype associations utilizing longitudinal EHR data. Previous PheWAS packages were developed mostly with smaller datasets and with earlier PheWAS approaches. PheTK was designed to simplify analysis and efficiently handle biobank-scale data. PheTK uses multithreading and supports a full PheWAS workflow including extraction of data from OMOP databases and Hail matrix tables as well as PheWAS analysis for both phecode version 1.2 and phecodeX. Benchmarking results showed PheTK took 64% less time than the R PheWAS package to complete the same workflow. PheTK can be run locally or on cloud platforms such as the All of Us Researcher Workbench (All of Us) or the UK Biobank (UKB) Research Analysis Platform (RAP).

AVAILABILITY AND IMPLEMENTATION

The PheTK package is freely available on the Python Package Index, on GitHub under GNU General Public License (GPL-3) at https://github.com/nhgritctran/PheTK, and on Zenodo, DOI 10.5281/zenodo.14217954, at https://doi.org/10.5281/zenodo.14217954. PheTK is implemented in Python and platform independent.

Clinical and genetic associations for night eating syndrome in a patient biobank

OBJECTIVE

Night eating syndrome (NES) is an eating disorder characterized by evening hyperphagia. Despite having a prevalence comparable to some other eating disorders, NES remains sparsely investigated and poorly characterized. The present study examined the phenotypic and genetic associations for NES in the clinical Mass General Brigham Biobank.

METHOD

Cases of NES were identified through relevant billing codes for eating disorders (F50.89/F50.9) and subsequent chart review; patients likely without NES were set as controls. Other diagnoses were determined from billing codes and collapsed into one of 1,857 distinct phenotypes based on clinical similarity. NES associations with diagnoses were systematically conducted in phenome-wide association scans using logistic regression models with adjustments for age, sex, race, and ethnicity. Polygenic scores for six related traits, namely for anorexia nervosa, depression, insomnia, sleep apnea, obesity, and type 2 diabetes were tested for associations with NES among participants of European ancestry using adjusted logistic regression models.

RESULTS

Phenome-wide scans comparing patients with NES against controls (cases n = 88; controls n = 64,539) identified associations with 159 clinical diagnoses spanning 13 broad disease groups including endocrine/metabolic and digestive diseases. Notable associations were evident for bariatric surgery, vitamin D deficiency, sleep disorders (sleep apnea, insomnia, and restless legs syndrome), and attention deficit hyperactivity disorder. The polygenic scores for insomnia and obesity were associated with higher odds of NES (insomnia: odds ratio [OR], 1.24; 95% CI, 1.07, 1.43; obesity: 1.98; 95% CI, 1.71, 2.28).

DISCUSSION

Complementary phenome-wide and genetic exploratory analyses provided information on unique and shared features of NES, offering insights that may facilitate its precise definition, diagnosis, and the development of targeted therapeutic interventions.

WebGWAS: A web server for instant GWAS on arbitrary phenotypes

Complex disease genetics is a key area of research for reducing disease and improving human health. Genome-wide association studies (GWAS) help in this research by identifying regions of the genome that contribute to complex disease risk. However, GWAS are computationally intensive and require access to individual-level genetic and health information, which presents concerns about privacy and imposes costs on researchers seeking to study complex diseases. Publicly released pan-biobank GWAS summary statistics provide immediate access to results for a subset of phenotypes, but they do not inform about all phenotypes or hand-crafted phenotype definitions, which are often more relevant to study. Here, we present WebGWAS, a new tool that allows researchers to obtain GWAS summary statistics for a phenotype of interest without needing access to individual-level genetic and phenotypic data. Our public web app can be used to study custom phenotype definitions, including inclusion and exclusion criteria, and to produce approximate GWAS summary statistics for that phenotype. WebGWAS computes approximate GWAS summary statistics very quickly (<10 seconds), and it does not store private health information. We also show how the statistical approximation underlying WebGWAS can be used to accelerate the computation of multi-phenotype GWAS among correlated phenotypes. Our tool provides a faster approach to GWAS for researchers interested in complex disease, providing approximate summary statistics in short order, without the need to collect, process, and produce GWAS results. Overall, this method advances complex disease research by facilitating more accessible and cost-effective genetic studies using large observational data.

GWAS highlights the neuronal contribution to multiple sclerosis susceptibility

Multiple Sclerosis (MS) is a chronic inflammatory and neurodegenerative disease affecting the brain and spinal cord. Genetic studies have identified many risk loci, that were thought to primarily impact immune cells and microglia. Here, we performed a multi-ancestry genome-wide association study with 20,831 MS and 729,220 control participants, identifying 236 susceptibility variants outside the Major Histocompatibility Complex, including four novel loci. We derived a polygenic score for MS and, optimized for European ancestry, it is informative for African-American and Latino participants. Integrating single-cell data from blood and brain tissue, we identified 76 genes affected by MS risk variants. Notably, while T cells showed the strongest enrichment, inhibitory neurons emerged as a key cell type, highlighting the importance of neuronal and glial dysfunction in MS susceptibility.

A One-Shot Lossless Algorithm for Cross-Cohort Learning in Mixed-Outcomes Analysis

In cross-cohort studies, integrating diverse datasets, such as electronic health records (EHRs), is both essential and challenging due to cohort-specific variations, distributed data storage, and data privacy concerns. Traditional methods often require data pooling or complex data harmonization, which can reduce efficiency and limit the scope of cross-cohort learning. We introduce mixWAS, a one-shot, lossless algorithm that efficiently integrates distributed EHR datasets via summary statistics. Unlike existing approaches, mixWAS preserves cohort-specific covariate associations and supports simultaneous mixed-outcome analyses. Simulations demonstrate that mixWAS outperforms conventional methods in accuracy and efficiency across various scenarios. Applied to EHR data from seven cohorts in the US, mixWAS identified 4,534 significant cross-cohort genetic associations among traits such as blood lipids, BMI, and circulatory diseases. Validation with an independent UK EHR dataset confirmed 97.7% of these associations, underscoring the algorithm's robustness. By enabling lossless cross-cohort integration, mixWAS improves the precision of multi-outcome analyses and expands the potential for actionable insights in healthcare research.

Oral Health, Inflammation, and Cardiometabolic Factors in the VA Million Veteran Program

INTRODUCTION

Associations between cardiometabolic comorbidities and self-reported oral health (OH) are often underexplored in large biobank datasets. While these associations are unaffected by dental care access, they could be mediated by immune responses and inflammation.

OBJECTIVES

This study assessed the associations between cardiometabolic comorbidities and self-reported OH, periodontitis, and tooth loss using the International Classification of Diseases (ICD) codes in participants from the U.S. Veterans Affairs Million Veteran Program (MVP), adjusting for immune and inflammatory covariates.

METHODS

Data from 154,167 MVP participants were extracted from January 2011 to September 2021, including lifetime cardiometabolic comorbidities, self-reported OH, ICD-coded periodontitis and tooth loss, and laboratory measurements. Multivariate logistic regression analysis was used to calculate the odds ratios of cardiometabolic comorbidities for self-reported OH, periodontitis, and tooth loss, adjusting for demographic, socioeconomic, cardiovascular, and inflammatory (neutrophil and lymphocyte cell counts) risk factors. A separate dataset was used for additional sensitivity analyses, adjusting for serum levels of C-reactive protein and albumin.

RESULTS

Complete data were analyzed for 154,167 participants (19%). Most participants (92%) were male and from European ancestry (94%). The mean age was 65.5 y (SD 11.4 y). Ten percent of participants had excellent self-reported OH. Fourteen percent had any periodontitis, and 17% had any tooth loss. Significant associations were found between tooth loss and congestive heart failure (odds ratio [OR], 1.74, P < 0.001) and peripheral vascular diseases (OR, 1.82, P < 0.001). There were also significant associations between congestive heart failure and self-reported OH (excellent versus "poor/fair/good/very good"), with increasing odds as self-reported OH declined (P < 0.001 for trend). These associations remained significant even after sensitivity analyses, albeit with slight attenuation.

CONCLUSION

This study of veterans underscores the important cardiometabolic links of self-reported poor OH and tooth loss, akin to those observed with periodontitis, even after adjusting for potential confounders related to demographics, lifestyle, and inflammation.

KNOWLEDGE TRANSFER STATEMENT

Exploring cardiometabolic associations with self-reported OH, clinically diagnosed periodontitis, and tooth loss using the ICD in the Veterans Affairs Million Veteran Program, we found significant associations. These associations persisted after adjustment for inflammatory confounders. These findings emphasized the benefit of assessing OH as a vital indicator of overall cardiometabolic health in large-scale biobank studies.

A Transcriptomic Comparative Study of Cranial Vasculature

In genetic studies of cerebrovascular diseases, the optimal vessels to use as controls remain unclear. Our goal is to compare the transcriptomic profiles among 3 different types of control vessels: superficial temporal artery (STA), middle cerebral arteries (MCA), and arteries from the circle of Willis obtained from autopsies (AU). We examined the transcriptomic profiles of STA, MCA, and AU using RNAseq. We also investigated the effects of using these control groups on the results of the comparisons between aneurysms and the control arteries. Our study showed that when comparing pathological cerebral arteries to control groups, all control groups presented similar responses in the activation of immunological processes, the regulation of intracellular signaling pathways, and extracellular matrix productions, despite their intrinsic biological differences. When compared to STA, AU exhibited upregulation of stress and apoptosis genes, whereas MCA showed upregulation of genes associated with tRNA/rRNA processing. Moreover, our results suggest that the matched case-control study design, which involves control STA samples collected from the same subjects of matched aneurysm samples in our study, can improve the identification of non-inherited disease-associated genes. Given the challenges associated with obtaining fresh intracranial arteries from healthy individuals, our study suggests that using MCA, AU, or paired STA samples as controls are feasible strategies for future large-scale studies investigating cerebral vasculopathies. However, the intrinsic differences of each type of control should be taken into consideration when interpreting the results. With the limitations of each control type, it may be most optimal to use multiple tissues as controls.

Illuminating the landscape of high-level clinical trial opportunities in the All of Us Research Program

OBJECTIVE

With its size and diversity, the All of Us Research Program has the potential to power and improve representation in clinical trials through ancillary studies like Nutrition for Precision Health. We sought to characterize high-level trial opportunities for the diverse participants and sponsors of future trial investment.

MATERIALS AND METHODS

We matched All of Us participants with available trials on ClinicalTrials.gov based on medical conditions, age, sex, and geographic location. Based on the number of matched trials, we (1) developed the Trial Opportunities Compass (TOC) to help sponsors assess trial investment portfolios, (2) characterized the landscape of trial opportunities in a phenome-wide association study (PheWAS), and (3) assessed the relationship between trial opportunities and social determinants of health (SDoH) to identify potential barriers to trial participation.

RESULTS

Our study included 181 529 All of Us participants and 18 634 trials. The TOC identified opportunities for portfolio investment and gaps in currently available trials across federal, industrial, and academic sponsors. PheWAS results revealed an emphasis on mental disorder-related trials, with anxiety disorder having the highest adjusted increase in the number of matched trials (59% [95% CI, 57-62]; P < 1e-300). Participants from certain communities underrepresented in biomedical research, including self-reported racial and ethnic minorities, had more matched trials after adjusting for other factors. Living in a nonmetropolitan area was associated with up to 13.1 times fewer matched trials.

DISCUSSION AND CONCLUSION

All of Us data are a valuable resource for identifying trial opportunities to inform trial portfolio planning. Characterizing these opportunities with consideration for SDoH can provide guidance on prioritizing the most pressing barriers to trial participation.

Clinical associations with treatment resistance in depression: An electronic health record study

Treatment resistance is common in major depressive disorder (MDD), yet clinical risk factors are not well understood. Using a discovery-replication design, we conducted phenome-wide association studies (PheWASs) of MDD treatment resistance in two electronic health record (EHR)-linked biobanks. The PheWAS included participants with an MDD diagnosis in the EHR and at least one antidepressant (AD) prescription. Participant lifetime diagnoses were mapped to phecodes. PheWASs were conducted for three treatment resistance outcomes based on AD prescription data: number of unique ADs prescribed, ≥1 and ≥2 CE switches. Of the 180 phecodes significantly associated with these outcomes in the discovery cohort (n = 12,558), 71 replicated (n = 8,206). In addition to identifying known clinical factors for treatment resistance in MDD, the total unique AD prescriptions was associated with additional clinical variables including irritable bowel syndrome, gastroesophageal reflux disease, symptomatic menopause, and spondylosis. We calculated polygenic risk of specific-associated conditions and tested their association with AD outcomes revealing that genetic risk for many of these conditions is also associated with the total unique AD prescriptions. The number of unique ADs prescribed, which is easily assessed in EHRs, provides a more nuanced measure of treatment resistance, and may facilitate future research and clinical application in this area.

Application of polygenic scores to a deeply phenotyped sample enriched for substance use disorders reveals extensive pleiotropy with psychiatric and somatic traits

Co-occurring psychiatric, medical, and substance use disorders (SUDs) are common, but the complex pathways leading to such comorbidities are poorly understood. A greater understanding of genetic influences on this phenomenon could inform precision medicine efforts. We used the Yale-Penn dataset, a cross-sectional sample enriched for individuals with SUDs, to examine pleiotropic effects of genetic liability for psychiatric and somatic traits. Participants completed an in-depth interview that provides information on demographics, environment, medical illnesses, and psychiatric and SUDs. Polygenic scores (PGS) for psychiatric disorders and somatic traits were calculated in European-ancestry (EUR; n = 5691) participants and, when discovery datasets were available, for African-ancestry (AFR; n = 4918) participants. Phenome-wide association studies (PheWAS) were then conducted. In AFR participants, the only PGS with significant associations was bipolar disorder (BD), all of which were with substance use phenotypes. In EUR participants, PGS for major depressive disorder (MDD), generalized anxiety disorder (GAD), post-traumatic stress disorder (PTSD), schizophrenia (SCZ), body mass index (BMI), coronary artery disease (CAD), and type 2 diabetes (T2D) all showed significant associations, the majority of which were with phenotypes in the substance use categories. For instance, PGSMDD was associated with over 200 phenotypes, 15 of which were depression-related (e.g., depression criterion count), 55 of which were other psychiatric phenotypes, and 126 of which were substance use phenotypes; and PGSBMI was associated with 138 phenotypes, 105 of which were substance related. Genetic liability for psychiatric and somatic traits is associated with numerous phenotypes across multiple categories, indicative of the broad genetic liability of these traits.

Examining genotype-phenotype associations of GRAM domain proteins using GWAS, PheWAS and literature review in cattle, human, pig, mouse and chicken

The GRAMD genes are involved in maintaining cholesterol homeostasis, apoptosis, cancer and production traits in livestock. A lipid-binding GRAM domain is implicated in lipid transport and metabolism. The functions of GRAMD proteins remain incompletely understood. The aim of the present study was therefore to investigate the associations between six GRAMD genes in cattle using data from the international genomic evaluation of the Interbull InterGenomics Centre and to evaluate genotype-phenotype associations in human, cattle, pig, mouse and, chicken. Genotyping of 55,013 bulls was performed using DNA microarrays and 11 SNPs were mapped to the five GRAMD genes. A phenome-wide association study (PheWAS) tested associations between the 11 SNPs and 36 traits. The integrated analysis of SNP effects, rankings, and clustering patterns revealed their potential for improving cattle productivity, health, and robustness, and established a baseline for the targeted improvement of cattle traits. This study lays the groundwork for functional experiments aimed at uncovering the mechanism of action of GRAMD genes and to evaluate the potential of using GRAMD sequence variants for selection programs in dairy cattle. The study presents an example of how the combination of GWAS and the PheWAS offers a promising toolbox for the systematic functional annotation of vertebrate genomes.

Whole-genome analysis of plasma fibrinogen reveals population-differentiated genetic regulators with putative liver roles

ABSTRACT

Genetic studies have identified numerous regions associated with plasma fibrinogen levels in Europeans, yet missing heritability and limited inclusion of non-Europeans necessitates further studies with improved power and sensitivity. Compared with array-based genotyping, whole-genome sequencing (WGS) data provide better coverage of the genome and better representation of non-European variants. To better understand the genetic landscape regulating plasma fibrinogen levels, we meta-analyzed WGS data from the National Heart, Lung, and Blood Institute's Trans-Omics for Precision Medicine (TOPMed) program (n = 32 572), with array-based genotype data from the Cohorts for Heart and Aging Research in Genomic Epidemiology Consortium (n = 131 340) imputed to the TOPMed or Haplotype Reference Consortium panel. We identified 18 loci that have not been identified in prior genetic studies of fibrinogen. Of these, 4 are driven by common variants of small effect with reported minor allele frequency (MAF) at least 10 percentage points higher in African populations. Three signals (SERPINA1, ZFP36L2, and TLR10) contain predicted deleterious missense variants. Two loci, SOCS3 and HPN, each harbor 2 conditionally distinct, noncoding variants. The gene region encoding the fibrinogen protein chain subunits (FGG;FGB;FGA) contains 7 distinct signals, including 1 novel signal driven by rs28577061, a variant common in African ancestry populations but extremely rare in Europeans (MAFAFR = 0.180; MAFEUR = 0.008). Through phenome-wide association studies in the VA Million Veteran Program, we found associations between fibrinogen polygenic risk scores and thrombotic and inflammatory disease phenotypes, including an association with gout. Our findings demonstrate the utility of WGS to augment genetic discovery in diverse populations and offer new insights for putative mechanisms of fibrinogen regulation.

Similar but Distinct Comorbidity Patterns Between Polycystic Ovary Syndrome and Endometriosis in Korean Women: A Nationwide Cohort Study

BACKGROUND

Polycystic ovary syndrome (PCOS) and endometriosis are widely recognized as significant risk factors affecting the reproductive health of women. The underlying mechanisms impacting fertility may vary, potentially leading to divergent outcomes. We aimed to examine and contrast the prevalence patterns of diseases coexisting with PCOS and endometriosis, using a large-scale nationwide insurance claims data from Asian women of reproductive age.

METHODS

We analyzed health insurance and examination data of 157,662 Korean women aged 15-45 years, drawn from the Korea National Health Insurance Service-National Sample Cohort database. International Classification of Disease, Tenth Revision codes were mapped to phenome-wide association study codes (phecodes). Subsequently, multivariate logistic regression was performed to assess the comorbidity patterns among patients diagnosed with PCOS and endometriosis and healthy control groups.

RESULTS

Our analysis revealed that PCOS was correlated with a wider range of metabolic disorders and symptoms, such as hyperlipidemia, type 2 diabetes, various gastrointestinal (GI) issues, and an array of pregnancy-related complications. Conversely, endometriosis was more prevalent among benign neoplasms of female reproductive and digestive organs, endometrial hyperplasia, and angina pectoris. Notably, infertility and glaucoma demonstrated significant associations with both conditions. Furthermore, a comparison of symptom-related codes in women with endometriosis revealed a predominance of pain-related symptoms, whereas those with PCOS exhibited a broader spectrum, encompassing pain, pruritus, GI problems, cough, fever, menstrual cycle disorders, edema, and dizziness.

CONCLUSION

PCOS and endometriosis, which are prevalent gynecological disorders affecting similar age groups of women, rarely co-occur and exhibit unique comorbidity profiles. Tailored healthcare strategies that take into account these distinct patterns have the potential to enhance long-term healthcare outcomes of affected patients. Further research is required to elucidate the underlying mechanisms and contrasting comorbidity profiles between PCOS and endometriosis.

Multisource representation learning for pediatric knowledge extraction from electronic health records

Electronic Health Record (EHR) systems are particularly valuable in pediatrics due to high barriers in clinical studies, but pediatric EHR data often suffer from low content density. Existing EHR code embeddings tailored for the general patient population fail to address the unique needs of pediatric patients. To bridge this gap, we introduce a transfer learning approach, MUltisource Graph Synthesis (MUGS), aimed at accurate knowledge extraction and relation detection in pediatric contexts. MUGS integrates graphical data from both pediatric and general EHR systems, along with hierarchical medical ontologies, to create embeddings that adaptively capture both the homogeneity and heterogeneity between hospital systems. These embeddings enable refined EHR feature engineering and nuanced patient profiling, proving particularly effective in identifying pediatric patients similar to specific profiles, with a focus on pulmonary hypertension (PH). MUGS embeddings, resistant to negative transfer, outperform other benchmark methods in multiple applications, advancing evidence-based pediatric research.

Development and validation of an electronic health record-based algorithm for identifying TBI in the VA: A VA Million Veteran Program study

The purpose of this study was to develop and validate an algorithm for identifying Veterans with a history of traumatic brain injury (TBI) in the Veterans Affairs (VA) electronic health record using VA Million Veteran Program (MVP) data. Manual chart review (n = 200) was first used to establish 'gold standard' diagnosis labels for TBI ('Yes TBI' vs. 'No TBI'). To develop our algorithm, we used PheCAP, a semi-supervised pipeline that relied on the chart review diagnosis labels to train and create a prediction model for TBI. Cross-validation was used to train and evaluate the proposed algorithm, 'TBI-PheCAP.' TBI-PheCAP performance was compared to existing TBI algorithms and phenotyping methods, and the final algorithm was run on all MVP participants (n = 702,740) to assign a predicted probability for TBI and a binary classification status choosing specificity = 90%. The TBI-PheCAP algorithm had an area under the receiver operating characteristic curve of 0.92, sensitivity of 84%, and positive predictive value (PPV) of 98% at specificity = 90%. TBI-PheCAP generally performed better than other classification methods, with equivalent or higher sensitivity and PPV than existing rules-based TBI algorithms and MVP TBI-related survey data. Given its strong classification metrics, the TBI-PheCAP algorithm is recommended for use in future population-based TBI research.

Development of electronic health record based algorithms to identify individuals with diabetic retinopathy

OBJECTIVES

To develop, validate, and implement algorithms to identify diabetic retinopathy (DR) cases and controls from electronic health care records (EHRs).

MATERIALS AND METHODS

We developed and validated electronic health record (EHR)-based algorithms to identify DR cases and individuals with type I or II diabetes without DR (controls) in 3 independent EHR systems: Vanderbilt University Medical Center Synthetic Derivative (VUMC), the VA Northeast Ohio Healthcare System (VANEOHS), and Massachusetts General Brigham (MGB). Cases were required to meet 1 of the following 3 criteria: (1) 2 or more dates with any DR ICD-9/10 code documented in the EHR, (2) at least one affirmative health-factor or EPIC code for DR along with an ICD9/10 code for DR on a different day, or (3) at least one ICD-9/10 code for any DR occurring within 24 hours of an ophthalmology examination. Criteria for controls included affirmative evidence for diabetes as well as an ophthalmology examination.

RESULTS

The algorithms, developed and evaluated in VUMC through manual chart review, resulted in a positive predictive value (PPV) of 0.93 for cases and negative predictive value (NPV) of 0.91 for controls. Implementation of algorithms yielded similar metrics in VANEOHS (PPV = 0.94; NPV = 0.86) and lower in MGB (PPV = 0.84; NPV = 0.76). In comparison, the algorithm for DR implemented in Phenome-wide association study (PheWAS) in VUMC yielded similar PPV (0.92) but substantially reduced NPV (0.48). Implementation of the algorithms to the Million Veteran Program identified over 62 000 DR cases with genetic data including 14 549 African Americans and 6209 Hispanics with DR.

CONCLUSIONS/DISCUSSION

We demonstrate the robustness of the algorithms at 3 separate healthcare centers, with a minimum PPV of 0.84 and substantially improved NPV than existing automated methods. We strongly encourage independent validation and incorporation of features unique to each EHR to enhance algorithm performance for DR cases and controls.