SEARCH · PubMed Health

Results for “Biobank”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 217 records · Page 12Linked to original sources

PubMed2026

Context-specific genetic effects inform endotypes and treatment in asthma.

BACKGROUND: Asthma has heterogeneous risk factors, subtypes, and treatments. It is often unclear how to stratify this heterogeneity in scientific studies and clinical care. Genetics could explain root causes of this clinical heterogeneity, called endotypes, but prior studies have used models that are not designed for complex diseases like asthma. OBJECTIVE: We aimed to find genetic effects that partly explain different asthma endotypes. METHODS: We used recent powerful and robust statistical models of context-specific genetic effects in complex traits. We identified genetic subtypes by clustering clinical asthma features in a case-control cohort, GALA II. We replicated the genetic endotypes in the UK Biobank with gene-context interaction tests. RESULTS: Asthma-associated single nucleotide polymorphisms, polygenic scores, and genome-wide heritability revealed subtype-specific genetic endotypes correlated with type 2 inflammation, allergy, and neuroticism. We validated the type 2 associations with molecular data including nasal RNA sequencing. In the UK Biobank, we replicated these endotypes and found they interact with several polygenic scores and drug-relevant genes. CONCLUSION: Our results show how context-specific genetic effects can unravel biomedically meaningful endotypes of complex disease and suggest novel precision treatment strategies.

Humans↗

PubMed2004

Hemoglobin A1c can be analyzed in blood kept frozen at -80 degrees C and is not commonly affected by hemolysis in the general population.

The ability of glycated hemoglobin A(1c) (HbA(1c)) to predict diabetes is unknown, but could be evaluated by analyses on samples stored in biobanks. The stability of HbA(1c) in long-term stored samples is, however, unknown. Moreover, the effect of hemolysis on HbA(1c) in the general population is not assessed. To explore these questions HbA(1c) was determined in 3 groups (n = 717) of samples with storage times at -80 degrees C differing between 10 years and 2 months. The results were compared with HbA(1c) analyzed in fresh blood samples (n = 174). The subjects were free from diabetes and aged 40 to 60 years. HbA(1c) was analyzed by cation exchange high-performance liquid chromatography (HPLC). The mean HbA(1c) results for the fresh and long-term stored samples were 4.25% +/- 0.39 and 4.19% +/- 0.43, respectively (P = not significant [NS]). The HbA(1c) levels in fresh and 2-month stored samples were essentially equal. There was no correlation between HbA(1c) in the fresh samples and the hemolysis parameters reticulocytes, haptoglobin, or bilirubin. HbA(1c) is apparently stable in samples long-term stored at -80 degrees C and is not commonly affected by hemolysis in the general population. HbA(1c) analyzed on biobank samples could be used to assess the predictive value for future diabetes and relationship to other morbidity and mortality.

Adult↗

PubMed2026

Genetic evidence for causality of late chronotype on metabolic syndrome in East Asians and Europeans.

CONTEXT: The impact of chronotype-defined as an individuals' inherent preference of sleep timing-and its genetic determinants on metabolic syndrome (MetS) has been less studied. OBJECTIVE: This study investigated the causal relationship between late chronotype and MetS using Mendelian randomization (MR) analysis, based on data from the Taiwan Biobank (TWB) and parallel analyses in the UK Biobank (UKB). METHODS: A total of 36,845 participants from TWB served as the discovery cohort, and 235,639 participants from UKB served as the replication cohort. Late chronotype was defined in TWB as a preference for bedtime after midnight, and in UKB as self-report as being an 'evening' person. The association between late chronotype and MetS, along with its components, was evaluated in TWB, and validated in UKB. Genome-wide association analyses for late chronotype were first conducted in TWB and then meta-analyzed with UKB. Polygenic risk scores (PRS) for late chronotype were constructed and tested for association with MetS. Causality between late chronotype and MetS was examined using one-sample MR analysis in TWB and validated in UKB. RESULTS: Late chronotype was significantly associated with MetS, as well as with central obesity, hyperglycemia, and hypertriglyceridemia, in both TWB and UKB (all P < 0.0083, considering Bonferroni correction). The constructed PRS of late chronotype also showed significant associations with MetS and several of its components (several P < 0.0083, considering Bonferroni correction). Findings from the one-sample MR analysis indicated a potential causal effect of late chronotype on MetS. CONCLUSIONS: This study provides evidence of a robust association between late chronotype and MetS across populations of diverse ancestry, including Taiwanese and European.

Humans↗

PubMed2026

Unravelling sex differences in the genetic architecture of anxiety.

BACKGROUND: Anxiety disorders show striking sex differences in prevalence, symptoms, and clinical characteristics, shaping how they manifest and are experienced. METHODS: Here, we report the first sex-specific meta-analysis of genome-wide association studies (GWAS) of anxiety, leveraging two of the largest biobank datasets, UK Biobank and All of Us, comprising 85,042 female cases with 196,789 controls and 36,732 male cases with 136,924 controls. Functional annotation, sex-specific polygenic scores (PGS), and genetic correlations were performed to assess genetic differences and functional implications. RESULTS: In females, 21 lead SNPs were significantly associated with anxiety, compared to five in males. Although the genetic correlation between sexes was high, it was significantly different from one, indicating partially distinct genetic architectures. In addition, both the SNP-based observed and liability-scale heritabilities (assuming a 2:1 female-to-male prevalence ratio) were significantly higher in females. Gene-based tests and functional prioritization identified different genes associated with anxiety in females and males. Moreover, genetic correlation analyses revealed stronger associations of female anxiety with attention-deficit/hyperactivity disorder (ADHD) and body mass index (BMI), whereas male anxiety showed stronger correlations with waist-hip-ratio-adjusted BMI. CONCLUSIONS: While the overall genetic architecture of anxiety is largely shared, our findings reveal distinct sex-specific genetic associations and correlations, highlighting the value of analyzing the sexes separately to uncover genetic signals that may be masked in sex-combined samples.

Female↗

PubMed2006

Wildlife biomaterial banking in Africa for now and the future.

The Wildlife Biological Resource Centre (wBRC) together with its partners in BioBank SA, have created a Biological Resource Bank (BRB) that is dedicated to the acquisition, processing, banking, using and provision of biomaterials to the scientific and conservation industry that are viable, diverse and representative of southern Africa's wildlife populations. Banked biomaterials include tissue such as muscle, kidney, fat, liver, embryos, fibroblast cultures, blood, sperm, hair, egg shells and other tissue, fluids and cells. Biomaterials are made available for research, biodiversity conservation and biotechnology development. Biomaterials are used in many disciplines, including genetics, reproduction, nutrition, and disease studies. Biomaterials from selected species are also useful for the detection and monitoring of Persistent Organic Pollutants and other potentially harmful substances found in the environment. Biomaterials are made available to third parties with prior consent from the biomaterials "owner" and only after the signing of a customised Material Transfer Agreement (MTA) or Cooperative Research and Development Agreement (CRADA). The training of staff from National and Provincial Game Reserves, Zoological Gardens, Animal Breeders and laboratories is carried out on a regular basis with the aim of securing good quality biomaterials. Sampling kits are made available to persons tasked with the collection of wildlife biomaterials. The Biobank SA consortium acts as an integrated resource centre linking partner collections. The consortium's operational arm, namely wBRC, is active in the development of relevant policy, regulations and legislation pertaining to biomaterials, including Access and Benefit Sharing systems. The main sponsor of the project is the Department of Science and Technology, National Government of South Africa.

Africa↗

PubMed2003

'Science is really needed--that's all I know': informed consent and the non-verbal practices of collecting blood for genetic research in northern Sweden.

In Vasterbotten County in northern Sweden a start-up biotech company has recently gained all commercial rights to one of the worlds largest population based research biobanks. The biobank and the company have publicly emphasized that all donors have given their informed consent to participate, but within the academy it has become debated whether people have been adequately informed. Based on anthropological fieldwork it is shown that many people do not read the information provided. The data do not, however, suggest that donors themselves perceive a lack of information. This article endeavours to make meaningful the apparent lack of interest among donors in the information they are offered. It is argued that the donation of blood should be analysed in its social and historical context rather than as a response to rational assessment of information of research purposes. It implies a conceptualisation of agency more aware of the intersubjective nature of moral negotiation than usually implied in studies of informed consent.

Biological Specimen Banks↗

PubMed2005

Increased level of linkage disequilibrium in rural compared with urban communities: a factor to consider in association-study design.

Few studies have investigated genetic differentiation within nonisolate European populations, despite the initiation of large national sample collections such as U.K. Biobank. Here, we used short tandem repeat markers to explore fine-scale genetic structure and to examine the extent of linkage disequilibrium (LD) within national subpopulations. We studied 955 unrelated individuals of local ancestry from nine Scottish rural regions and the urban center of Edinburgh, as well as 96 unrelated individuals from the general U.K. population. Despite little overall differentiation on the basis of allele frequencies, there were clear differences among subpopulations in the extent of pairwise LD, measured between a subset of X-linked markers, that reflected presumed differences in the depths of the underlying genealogies within these subpopulations. Therefore, there are strategic advantages in studying rural subpopulations, in terms of increased power and reduced cost, that are lost by sampling across regions or within urban populations. Similar rural-urban contrasts are likely to exist in many other populations with stable rural subpopulations, which could influence the design of genetic association studies and national biobank data collections.

Genetics, Population↗

PubMed2026

STABIX: summary-statistic-based GWAS indexing and compression.

MOTIVATION: Genome-wide association studies (GWAS) are widely used to investigate the role of genetics in disease traits, but the resulting file sizes from these studies are large, posing barriers to efficient storage, sharing, and querying. This issue is especially important for biobanks like the UK Biobank that publish GWAS for thousands of traits, increasing the volume of data that must be effectively managed. Current compression and query methods reduce file sizes and allow for quick genomic position-based queries but do not provide utility for quickly finding loci based on their summary statistics. For example, finding all SNVs in a particular p-value range would require decompressing and scanning the whole file. We propose a new tool, STABIX, which introduces summary-statistic-based queries and improves upon the standard bgzip compression and Tabix query tool in both compression ratio and decompression speed. RESULTS: When applied to 10 GWAS files from PanUKBB, STABIX created smaller compressed data and indices than Tabix for all files, where bgzip and tbi files were an average of 1.2 times the size of STABIX compressed files and indexes. In the same 10 files, STABIX per gene decompression was, on average 7× faster than Tabix per gene decompression, and achieved faster per gene decompression times for over 99% of nearly 20,000 genes. AVAILABILITY AND IMPLEMENTATION: Software freely available for download at GitHub: https://github.com/kristen-schneider/stabix/.

Genome-Wide Association Study↗

PubMed2025

Sparse polygenic risk score inference with the spike-and-slab LASSO.

MOTIVATION: Large-scale biobanks, with rich phenotypic and genomic data across hundreds of thousands of samples, provide ample opportunities to elucidate the genetics of complex traits and diseases. Consequently, there is growing demand for robust and scalable methods for disease risk prediction from genotype data. Inference in this setting is challenging due to the high-dimensionality of genomic data, especially when coupled with smaller sample sizes. Popular Polygenic Risk Score (PRS) inference methods address this challenge by adopting sparse Bayesian priors or penalized regression techniques, such as the Least Absolute Shrinkage and Selection Operator (LASSO). However, the former class of methods are not as scalable and do not produce exact sparsity, while the latter tends to over-shrink large coefficients. RESULTS: In this study, we present SSLPRS, a novel PRS method based on the Spike-and-Slab LASSO (SSL) prior, which offers a theoretical bridge between the two frameworks. We extend previous work to derive a coordinate-ascent inference algorithm that operates on GWAS summary statistics, which is orders-of-magnitude more efficient than corresponding individual-level-based implementations. To illustrate the statistical properties of the proposed model, we conducted experiments involving nine simulation configurations and nine quantitative phenotypes from the UK Biobank. Our results demonstrate that SSLPRS is competitive with state-of-the-art methods in terms of prediction accuracy and exhibits superior variable selection performance, especially in sparse genetic architectures. In simulations, this translates to upwards of 50% improvement in positive predictive value. In analysis of real phenotypes, we show that selected variants are highly enriched for meaningful genomic annotations and have better replication rates in larger meta-analyses. AVAILABILITY AND IMPLEMENTATION: SSLPRS is available in the open-source package https://github.com/li-lab-mcgill/penprs.

Multifactorial Inheritance↗

PubMed2026

A novel method for across-chromosome phasing without relative data.

MOTIVATION: Across-chromosome phasing identifies which haplotypes of different chromosomes come from the same parent. This differs from within-chromosome phasing, which uses linkage disequilibrium patterns to determine which alleles were co-inherited within each chromosome but does not match haplotypes across different chromosomes. While across-chromosome phasing can be conducted using genotypes from parents or close relatives, current methods perform poorly for samples of unrelated individuals. Here, we introduce a novel approach for across-chromosome phasing that employs a window-based SNP-similarity metric, eliminating the need for data from close relatives or detection of identical-by-descent haplotypes. RESULTS: Using UK Biobank offspring with both parents genotyped as a gold standard, we evaluated the performance of our method by phasing the offspring without using parental data. In genomic data with no within-chromosome phase errors, our algorithm achieved a mean across-chromosome phasing accuracy of 95%, with 53% of individuals phased perfectly. When data was pre-phased computationally using a standard within-chromosome phasing algorithm, mean accuracy for across-chromosome phasing dropped to 83.1%. Thus, our method is limited primarily by the accuracy of within-chromosome phasing accuracy and can approach near-perfect across-chromosome phasing accuracy as within-chromosome phasing accuracy improves. AVAILABILITY AND IMPLEMENTATION: The implementation was executed within a multi-node computational environment of University of Colorado Boulder Research Computing (Blanca Cluster: https://www.colorado.edu/rc/resources/blanca), employing parallelization techniques in the C programming language. The source code has been made publicly accessible online at https://github.com/emmanuelsapin/AcrossChromosomesPhasing, thereby facilitating reproducibility of the results for researchers with authorized access to the UK Biobank dataset.

Algorithms↗

PubMed2022

Genetically predicted on-statin LDL response is associated with higher intracerebral haemorrhage risk.

Statins lower low-density lipoprotein cholesterol and are widely used for the prevention of atherosclerotic cardiovascular disease. Whether statin-induced low-density lipoprotein reduction increases risk of intracerebral haemorrhage has been debated for almost two decades. Here, we explored whether genetically predicted on-statin low-density lipoprotein response is associated with intracerebral haemorrhage risk using Mendelian randomization. Using genomic data from randomized trials, we derived a polygenic score from 35 single nucleotide polymorphisms of on-statin low-density lipoprotein response and tested it in the population-based UK Biobank. We extracted statin drug and dose information from primary care data on a subset of 225 195 UK Biobank participants covering a period of 29 years. We validated the effects of the genetic score on longitudinal low-density lipoprotein measurements with generalized mixed models and explored associations with incident intracerebral haemorrhage using Cox regression analysis. Statins were prescribed at least once to 75 973 (31%) of the study participants (mean 57 years, 55% females). Among statin users, mean low-density lipoprotein decreased by 3.45 mg/dl per year [95% confidence interval (CI): (-3.47, -3.42)] over follow-up. A higher genetic score of statin response [1 standard deviation (SD) increment] was associated with significant additional reductions in low-density lipoprotein levels [-0.05 mg/dl per year, (-0.07, -0.02)], showed concordant lipidomic effects on other lipid traits as statin use and was associated with a lower risk for incident myocardial infarction [hazard ratio per SD increment 0.98 95% CI (0.96, 0.99)] and peripheral artery disease [hazard ratio per SD increment 0.93 95% CI (0.87, 0.99)]. Over a 11-year follow-up period, a higher genetically predicted statin response among statin users was associated with higher intracerebral haemorrhage risk in a model adjusting for statin dose [hazard ratio per SD increment 1.16, 95% CI (1.05, 1.28)]. On the contrary, there was no association with intracerebral haemorrhage risk among statin non-users (P = 0.89). These results provide further support for the hypothesis that statin-induced low-density lipoprotein reduction may be causally associated with intracerebral haemorrhage risk. While the net benefit of statins for preventing vascular disease is well-established, these results provide insights about the personalized response to statin intake and the role of pharmacological low-density lipoprotein lowering in the pathogenesis of intracerebral haemorrhage.

Cerebral Hemorrhage↗

PubMed2026

Sex-specific associations of the plasma-proteome with incident coronary artery disease.

AIMS: The etiology of coronary artery Disease (CAD) appears different for men and women, yet insights into underlying sex-specific biological mechanisms are limited. We integrated genomic and proteomic analyses to investigate sex-specific associations of the plasma-proteome with CAD. METHODS AND RESULTS: In 40,829 UK Biobank participants (free-of-CAD, baseline-365 days thereafter; 55% women; mean age 56.9 ± 8.1 years), we examined associations between 2,922 plasma proteins and incident CAD over a median follow-up of 13.7 years (IQR 13.1-14.4) using multivariable-adjusted Cox proportional hazards models. Sex-specific analyses identified 440 female exclusive and 32 male exclusive proteins associated with incident CAD (FDR-corrected p < 0.05), revealing distinct pathway enrichments, including innate immune response in women and angiogenesis in men. Causality was assessed through combined and sex-stratified two-sample Mendelian randomization (MR) using inverse-variance-weighted analyses with genome wide association summary statistics from 422,108 men (61,969 cases) and 521,695 women (27,128 cases) (UK Biobank, FinnGen freeze 9). Integration of direct sex-protein interaction analyses with sex-combined MR identified 59 proteins with evidence for sex-specific causal effects. Four proteins demonstrated concordant directionality in sex-stratified MR analyses (n = 943,803) and multivariable regression models, namely CDKN2D, MYH9, and SKAP2 (women), and CTSH (men). To assess translational relevance, prioritized targets were further evaluated in secondary major adverse cardiovascular events among carotid endarterectomy patients (MACE; Athero-Express) and acute myocardial infarction (AMI; MISSION!) using plasma proteomics and ELISA. After further top-target identification in the context of MACE and AMI, clinical drug candidates were identified through a machine learning framework, including CTSH (men), and TNFRSF4 (both sexes). CONCLUSIONS: We identified sex-specific associations of proteins and biological pathways with incident CAD. Whereas the majority of proteins had consistent associations in both men and women, our findings suggest a degree of sex-specific pathogenesis with evidence for potential causality, opening new alleys for tailored prevention strategies and clinical cardiovascular risk management.

Journal Article↗

PubMed2026

Multi-Omics Genome-Wide to Explore the Formation and Development Targets for Intracranial Aneurysms.

Intracranial aneurysms (IAs) represent a significant and potentially life-threatening category of disease, and there is currently a lack of effective treatment options aimed at preventing the progression of the disease. Accordingly, this study is dedicated to exploring and identifying effective drug targets that can help in the prevention of both the formation and rupture of IAs, along with a detailed examination of the underlying potential mechanisms involved in these processes. The data related to IAs for this research was obtained from the ISGC Biobank and UK Biobank. Then, we investigated the possible biological functions and unintended consequences of targeting the specific genes that were highlighted in IAs by using mediation analysis, virtual knockout experiments, and PW-MR studies. A total of 5 unique potential drug targets for IAs (FKTN, MAP3K1, PSMA4, SLC22A4, ADAM17), 4 unique potential drug targets for SAH (PSMA4, ADAM17, GPR160, SLC22A4), and 2 unique potential drug targets for UIA (SLC22A4, PRCP) were identified across brain or blood samples. Among the various candidates identified, SLC22A4 has emerged as a promising potential drug target, showing significant expression levels in both blood and brain tissues. Additionally, phenome-wide MR of SLC22A4 across 32 selected phenotypes did not identify statistically significant adverse associations after FDR correction. Virtual knockout (KO) experiments on SLC22A4 revealed that SLC22A4 KO disrupted 81 genes, all of which are involved in IAs-related pathways. Besides, we recognized BRD-K85337334 as potential candidates for targeting SLC22A4. This research indicates that an increase in SLC22A4 gene expression within the blood or brain is directly linked to a heightened risk of IAs rupture, which will aid in prioritizing the development of drugs for IAs.

Humans↗

PubMed2026

Investigating the causal effects of physical activity and sedentary behavior on hernia risk: A two-sample Mendelian randomization approach.

The global prevalence of hernia is increasing, particularly due to the aging population. Although physical activity and sedentary behavior are known to influence various health outcomes, their specific roles in hernia development remain inadequately investigated. This study aimed to systematically assess the causal relationships between physical activity, sedentary behavior, and the risk of hernia development using Mendelian randomization (MR) analysis. We used genome-wide association study data from the UK Biobank and FinnGen Biobank to explore the causal effects of sedentary behavior on hernia risk via 5 distinct MR approaches. To ensure the reliability of our risk models, we performed sensitivity analyses, including Cochran's Q test, MR-Egger intercept analysis, leave-one-out analysis, and funnel plots. Furthermore, we employed multivariable MR analysis to determine the independent causal effects of both physical activity and sedentary behavior. Our findings indicate that moderate-to-vigorous physical activity (MVPA) was significantly associated with an increased risk of inguinal hernia, with an odds ratio of 1.844 (95% confidence interval, 1.181-2.879; P = .007). Associations between sedentary behavior and diaphragmatic or umbilical hernia were inconsistent across methods and were sensitive to pleiotropy; thus, no robust causal relationship was established. Multivariable MR analysis further confirmed the independent and significant causal relationship between MVPA and inguinal hernia, with an odds ratio of 1.901 (95% confidence interval, 1.098-3.291; P = .022). Our study highlights a significant association between MVPA and an increased risk of inguinal hernia, suggesting that physical activity should be carefully considered in preventive strategies for hernias. However, the relationship between sedentary behavior and hernia development warrants further investigation.

Mendelian Randomization Analysis↗

PubMed2025

Phenotypic and Genetic Associations Between Cardiovascular Disease Subtypes and Alzheimer's Disease.

BACKGROUND: Cardiovascular disease (CVD) and Alzheimer's disease (AD) are major public health concerns that share overlapping risk factors and potential mechanistic pathways. While vascular contributions to cognitive decline are well-documented, the specific relationships between AD and different CVD subtypes remain poorly understood. METHODS: We examined associations between AD and 11 CVD subtypes using logistic regression models in two large biobanks: the UK Biobank (n = 502,133) and the All of Us Research Program (n = 287,011). Models were adjusted for demographic, lifestyle, and clinical covariates. We also explored genetic overlap between AD and CVD traits through colocalization of significant single nucleotide polymorphisms (SNPs) (p < 5×10-8) using genome-wide association study (GWAS) data. RESULTS: Most CVD subtypes were significantly associated with AD in both cohorts. Hypotension had the strongest and most consistent association, followed by hypertension and cerebral infarction. Acute myocardial infarction was the only subtype not significantly linked to AD. Genetic analyses revealed shared loci between AD and CVD-related traits, particularly in regions near APOE, MAPT, and genes influencing myocardial structure and vascular function. CONCLUSIONS: This study identifies subtype-specific CVD associations with AD across two diverse cohorts and highlights shared genetic architecture underlying heart-brain interactions. These findings underscore the importance of vascular health in AD risk and suggest that certain CVD subtypes, especially hypotension, may play underrecognized roles in cognitive decline.

Alzheimer’s disease↗

PubMed2025

The impact of polygenic score and socioeconomic status in predicting risk for 19 complex diseases.

Both socioeconomic circumstances and genetic predisposition shape disease risk, yet their joint contribution across diseases has not been systematically examined. We studied 19 high-burden diseases in 743,194 participants (729,928 European; 13,266 non-European ancestry) from FinnGen, the UK Biobank, and Generation Scotland. Higher educational attainment was associated with lower risk of most conditions, but with higher risk of most common cancers. These associations were largely independent of disease-specific polygenic scores (PGSs). For seven out of 19 diseases, PGSs showed stronger effects among individuals with high education. Joint inclusion of education and PGSs modestly improved prediction for 14 and 10 out of 19 diseases in FinnGen and the UK Biobank, respectively. PGS associations were consistent across ancestries, whereas education effects were less stable; results using an alternative socioeconomic measure were directionally similar but smaller. Our findings highlight the distinct and partly interacting contributions of socioeconomic and genetic factors to disease risk.

Journal Article↗

PubMed2025

Privacy-Enhancing Sequential Learning under Heterogeneous Selection Bias in Multi-Site EHR Data.

OBJECTIVE: To develop privacy-enhancing statistical methods for estimation of binary disease risk model association parameters across multiple electronic health record (EHR) sites with heterogeneous selection mechanisms, without sharing raw individual-level data. We illustrate their utility through a cross-biobank analysis of smoking and 97 cancer subtypes using data from the NIH All of Us (AOU) and the Michigan Genomics Initiative (MGI). MATERIALS AND METHODS: Large-scale biobanks often follow heterogeneous recruitment strategies and store data in separate cloud-based platforms, making centralized algorithms infeasible. To address this, we propose two decentralized sequential estimators namely, Sequential Pseudo-likelihood (SPL) and Sequential Augmented Inverse Probability Weighting (SAIPW) that leverage external population-level information to adjust for selection bias, with valid variance estimation. SAIPW additionally protects against misspecification of the selection model using flexible machine learning based auxiliary outcome models. We compare SPL and SAIPW with the existing Sequential Unweighted (SUW) estimator and with centralized and meta learning extensions of IPW and AIPW in simulations under both correctly specified and misspecified selection mechanisms. We apply the methods to harmonized data from MGI ( n = 50,935) and AOU ( n = 241,563) to estimate smoking-cancer associations. RESULTS: In simulations, SUW exhibited substantial bias and poor coverage. SPL and SAIPW yielded unbiased estimates with valid coverage probabilities under correct model specification, with SAIPW remaining robust under selection model misspecification. Both approaches showed no notable efficiency loss relative to centralized methods. Meta-learning methods were efficient for large sites but failed in settings with small cohort sizes and rare outcome prevalence. In real-data analysis, strong associations were consistently identified between smoking and cancers of the lung, bladder, and larynx, aligning with established epidemiological evidence. CONCLUSION: Our framework enables valid, privacy-enhancing inference across EHR cohorts with heterogeneous selection, supporting scalable, decentralized research using real-world data.

Journal Article↗

PubMed2006

High-potentiality preliminary selection criteria and transformation time-dependent factors analysis for establishing Epstein-Barr virus transformed human lymphoblastoid cell lines.

Infection of freshly isolated and cryopreserved lymphocytes with Epstein-Barr virus (EBV) leads to the establishment of human B lymphoblastoid cell lines (LCLs). Techniques for optimal infection of the lymphocytes are vital for the establishment of a human biobank. The present study found that more than half (58-86%) of such established LCLs had transport times of less than 48 h, cell densities exceeding 10(6) cells/ml and cell viabilities greater than 90%. After EBV infection, 3306 freshly isolated lymphocytes required 30.0 +/- 0.1 days to become LCLs. Conversely, 1210 cryopreserved lymphocytes required 36.2 +/- 0.4 days. Cell density and viability of the culture affected transformation time in freshly isolated lymphocytes. On the other hand, blood transport time, cryopreservation time and initial cell viability were major factors in cryopreserved specimens. These results contribute to general information concerning the establishment of a human biobank for EBV infected cells.

Animals↗