ZSZhuozheng Shi

Zhuozheng Shi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Zhuozheng Shi has a total of 48 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A unified framework for local-ancestry-aware genetic association analysis across biobanks

    by Linfeng Hu, Taotao Tan, Kai Yuan, Ying Wang, Bram L. Gorissen, Yi-Sian Lin, Pragati Kore, Wenhan Lu, Ravi Mandla, Zhuozheng Shi, et al.

    Biobanks increasingly include individuals with admixed genomes, yet conventional genome-wide association study frameworks either exclude participants who cannot be confidently assigned to a discrete ancestry group or ignore ancestry-specific effects. We present FELIX, a scalable framework for local-ancestry-aware genetic analysis that retains all participants without requiring discrete ancestry assignment. FELIX combines a compact ancestry-resolved genotype representation (FELIXla) with an adaptive association test that jointly evaluates shared-effect and ancestry-specific models at each variant (FELIXassoc). Simulations demonstrated well-calibrated inference under case-control imbalance and power that adapted to the locus-optimal model. Across 24 phenotypes in 240,038 All of Us participants, FELIX analyzed the 12.1% of individuals excluded by global-ancestry clustering and identified 15.4% more genome-wide significant loci than global-ancestry meta-analysis. Additional discoveries arose from recovering ancestry-specific haplotypes carried by admixed participants and from detecting ancestry-dependent marginal effects. Full-cohort effect estimates also improved polygenic score prediction across ancestries and traits.

  • Open Access

    CalPred yields calibrated intervals for polygenic risk prediction

    by Zhuozheng Shi, Zixuan (Eleanor) Zhang, Ravi Mandla, Kangcheng Hou, Bogdan Pasaniuc

    Polygenic scores (PGS) have emerged as a useful biomarker for stratification of high-risk individuals in genomic medicine, with prediction intervals arising as a principled approach to incorporate statistical uncertainty in their individual-level predictions. In contrast to recent reports by Xu et al 7 , we show that CalPred 6 provides well-calibrated prediction intervals that contain the trait phenotypes at targeted confidence levels. CalPred maintains calibration when PGS performance varies across contextual factors (e.g., ancestry, age, sex, or socio-economic factors) whereas PredInterval 7 – a recently introduced method that focuses on marginal calibration across all individuals – exhibits miscalibration.


  • Open Access

    Implementation of the genome-informed risk assessment (GIRA) may lead to large disruptions to the health system

    by Sandra Lapinska, Xinzhe Li, Ravi Mandla, Zhuozheng Shi, Veronica Tozzo, Alexander Flynn-Carroll, Marylyn D Ritchie, Daniel J Rader, Bogdan Pasaniuc

    The Genome Informed Risk Assessment (GIRA) report from eMERGE has become a standard approach to implement genomic precision medicine at scale. Here, we assess GIRA's utility and impact in a health care system independent of eMERGE, focusing on 9 adult conditions using the Penn Medicine Biobank (PMBB, n=48,279). We find a large number of patients - 50.1% (n=24,185) - were deemed by GIRA as high-risk for at least one of the 9 conditions with 30.4% (n=14,676) due to polygenic and/or monogenic risk. Stratifying by ancestry revealed significant differences in high-risk proportions, with higher rates in African/African American (AFR) (56.6% vs. 50.1%, p=7.43x10-36) and lower rates in East (42.0%) and South Asian (40.0%). Increased high-risk rates were observed in the lowest quartile of social deprivation index, highlighting the influence of environmental factors and access to care on GIRA's utility. GIRA was a good predictor of prevalent cases (in-line with the eMERGE GIRA reported results); incident case prediction was substantially attenuated for 5 of the 9 conditions (e.g., OR of 2.36 vs. HR of 1.31 for atrial fibrillation (AFIB)). We find demographic compositions of high-risk patients differed from the incident cases for some of the conditions; for example, high-risk for AFIB individuals where enriched for European ancestries in contrast with incident AFIB cases that were enriched for AFR ancestries. Overall, our results show the accuracy of GIRA as a biomarker to stratify high-risk patients for precision medicine and highlight implementation challenges in its impact on the health system if implemented at scale.

  • Open Access

    Statistical uncertainty explains the poor agreement in polygenic scoring for type 2 diabetes

    by Ravi Mandla, Xinzhe Li, Zhuozheng Shi, Sarah A. Abramowitz, Sandra Lapinska, Penn Medicine Biobank, Michael G. Levin, Scott M. Damrauer, Bogdan Pasaniuc

    Polygenic scores (PGS) have emerged as an important tool for genetic risk prediction in medicine to identify individuals at high-risk for disease. A major limitation in their implementation is the apparent disagreement among scores for the same individual decreasing their interpretability and utility in clinical settings. Here we show that the poor agreement across PGSes for type 2 diabetes (T2D) is fully explained by statistical uncertainty in PGS-based prediction; individual-level uncertainty estimates from a single PGS explain the variability across existing PGSes. We provide an approach for the selection of high-risk individuals that incorporates measures of uncertainty and show that individuals with high confidence based on their PGS uncertainty have higher risk agreement across existing PGS and are more likely to develop T2D than high-risk individuals based on only point estimates of PGS. Together, these findings shed light on the factors underlying a roadblock in PGS implementation and underscore the need to incorporate uncertainty in PGS-based predictions.

  • Open Access

    Genome-wide analysis implicates inner ear development in Ménière’s disease

    by Zhuozheng Shi, Ravi Mandla, Jingjing Li, Xinzhe Li, Zixuan (Eleanor) Zhang, Sixing Chen, Sandra Lapinska, Alexander O. Flynn-Carroll, Bogdan Pasaniuc, Douglas J. Epstein, et al.

    Ménière’s disease (MD) is a chronic inner ear disorder characterized by recurrent vertigo, fluctuating sensorineural hearing loss, and tinnitus. Despite these distinctive symptoms, its etiology remains poorly understood. We performed a genome-wide meta-analysis of 8,969 cases and 1,962,542 controls across five large biobanks, identifying five independent genome-wide significant loci and estimating an observed-scale SNP heritability of 7% (SE 0.8%), consistent with a modest but significant genetic contribution to MD risk. Fine-mapping and integrative functional analyses implicate two convergent biological processes – developmental regulation of the inner ear, involving EYA4, EYA1 , and LMO4 – and retinoic acid metabolism, with loci near CYP26A1/C1 and ALDH1A2 suggesting disrupted RA signaling in sensory and fluid-pressure homeostasis. These developmental regulator genes are robustly expressed in fetal and adult human inner ear cell types, supporting a model in which altered developmental programs predispose to adult vestibular and auditory dysfunction. Phenome-wide and genetic correlation analyses further reveal shared genetic architecture between MD and related traits, including vertigo, tinnitus, hearing loss, migraine, and sleep apnea, situating MD within a broader spectrum of sensory and neurological disorders. Collectively, these findings establish a genetic framework for Ménière’s disease risk and implicate developmental regulators and retinoic acid signaling as key contributing pathways.

  • Open Access

    A biobank-scale test of marginal epistasis reveals genome-wide signals of polygenic interaction effects

    by Boyang Fu, Ali Pazokitoroudi, Zhuozheng Shi, Asha Kar, Albert Xue, Aakarsh Anand, Prateek Anand, Zhengtong Liu, Richard Border, Päivi Pajukanta, et al.

    The contribution of genetic interactions (epistasis) to human complex trait variation remains poorly understood due, in part, to the statistical and computational challenges involved in testing for interaction effects. Here we introduce FAME (FAst Marginal Epistasis test), a method that can test for marginal epistasis of a single-nucleotide polymorphism (SNP) on a quantitative trait (whether the effect of an SNP on the trait is modulated by genetic background). FAME is computationally efficient, enabling tests of marginal epistasis on biobank-scale data. Applying FAME to genome-wide association study (GWAS)-significant trait-SNP associations across 53 quantitative traits and ≈300 000 unrelated White British individuals in the UK Biobank (UKBB), we identified 16 significant marginal epistasis signals across 12 traits ( $$P < \frac{5\times {10}^{-8}}{53}$$ P < 5 × 10 − 8 53 ). Leveraging the scalability of FAME, we further localized marginal epistasis signals across chromosomes and estimated the proportion of variance explained by marginal epistasis effects. Our study provides evidence for interactions between individual genetic variants and polygenic background influencing complex traits.


  • Open Access

    Hidden structure in polygenic scores and the challenge of disentangling ancestry interactions in admixed populations

    by Alan J. Aw, Ravi Mandla, Zhuozheng Shi, Penn Medicine Biobank, Bogdan Pasaniuc, Iain Mathieson

    The extent to which genetic effects vary across ancestries remains a central question in human genetics, with direct implications for improving polygenic prediction. Recent studies have found that average causal effects are similar across ancestries, but empirical evidence and theoretical considerations suggest that interactions may still contribute to poor portability of polygenic scores. Here, we leverage the ancestry mosaicism of admixed individuals to develop models of genetic interactions with local and global ancestry. We show that these models capture epistasis but are also consistent with highly similar average causal effects, clarifying that the latter observation does not exclude the possibility of gene-ancestry interaction. Next, we investigate how models of local and global interactions differ in polygenic prediction. Focusing on continuous traits, we show that if causal variants are known and causal effects differ across ancestries, then the global and local models can be differentiated using partial polygenic scores — scores computed on ancestry-specific genomic segments — while standard polygenic scores remain indistinguishable under both models. However, if causal variants are unknown, differences in linkage disequilibrium (LD) patterns confound this analysis, and the models are not practically distinguishable without knowledge of the LD between causal and tagging variants. Our findings motivate future developments of model-sensitive strategies for individual-level genetic risk prediction.