Jingshu Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Jingshu Wang has a total of 72 co-authors. Here are the top co-authors by number of shared publications:
- Qingyuan Zhao
Statistical Laboratory, University of Cambridge
8 shared publications - Nancy Ruonan Zhang
Department of Statistics, The Wharton School, University of Pennsylvania, Philadelphia, USA
5 shared publications - Jack Bowden
Department of Clinical and Biomedical Sciences, Faculty of Health and Life Sciences, University of Exeter
4 shared publications - Zixuan Wu
Department of Statistics, University of Chicago, Chicago, IL 60637, USA
4 shared publications - Dylan S. Small
Department of Statistics and Data Science, University of Pennsylvania , 265 South 37th Street , Philadelphia, Pennsylvania 19104
3 shared publications - Gibran Hemani
University of Bristol
3 shared publications
Latest Publications
- 27 October 2026
- Published by Taylor & Francis in Journal of the American Statistical Association
- p. 1-61
Sensitivity Analysis for the Test-Negative Design
- 27 October 2026
- Published by Taylor & Francis in Journal of the American Statistical Association
- p. 1-61
Sensitivity Analysis for the Test-Negative Design
- Open Access
Transcriptional perturbation effects remain structured beyond the leading global trend
Prior work has described a global transcriptional response to gene perturbation, with many perturbations moving cells along a common direction. We find that perturbations contribute differently to this leading global trend and that substantial organization remains beyond it, including transferable perturbation-specific effects across datasets and a common hierarchy of responsive genes across perturbations. These patterns make simple strategies surprisingly competitive even in predicting effects beyond the leading global component.
Open AccessTranscriptional perturbation effects remain structured beyond the leading global trend
Prior work has described a global transcriptional response to gene perturbation, with many perturbations moving cells along a common direction. We find that perturbations contribute differently to this leading global trend and that substantial organization remains beyond it, including transferable perturbation-specific effects across datasets and a common hierarchy of responsive genes across perturbations. These patterns make simple strategies surprisingly competitive even in predicting effects beyond the leading global component.
- Open Access
Improving estimation efficiencies for family-based GWAS by integrating large external data
Family-based genome-wide association studies (GWAS) can separate direct from indirect genetic effects such as genetic nurture, population stratification, and assortative mating, yet these designs often suffer from limited statistical power because large samples of genotyped trios and sibships are rare. We introduce a calibration framework that improves the efficiency of within-family GWAS by integrating three readily available summary statistics for each SNP: the within-family association, the corresponding population-based estimate from the same family sample, and an external population-based estimate from a large GWAS. The method does not require individual-level data and is compatible with generalized linear models used for both continuous and binary traits. Theoretical results show that calibration can reduce variance by up to fifty percent in trio designs and up to twenty-five percent in sibling designs, equivalent to doubling the effective sample size for trios. Simulations confirm the accuracy and unbiasedness of the calibrated estimator, and applications to UK Biobank family data demonstrate substantial precision gains and improved downstream Mendelian Randomization inference. Analysis of published within-sibship GWAS summary statistics further illustrates that the approach can be applied directly to publicly available data. Together, these results show that calibration provides a practical and powerful way to enhance family-based genetic analyses.
Open AccessImproving estimation efficiencies for family-based GWAS by integrating large external data
Family-based genome-wide association studies (GWAS) can separate direct from indirect genetic effects such as genetic nurture, population stratification, and assortative mating, yet these designs often suffer from limited statistical power because large samples of genotyped trios and sibships are rare. We introduce a calibration framework that improves the efficiency of within-family GWAS by integrating three readily available summary statistics for each SNP: the within-family association, the corresponding population-based estimate from the same family sample, and an external population-based estimate from a large GWAS. The method does not require individual-level data and is compatible with generalized linear models used for both continuous and binary traits. Theoretical results show that calibration can reduce variance by up to fifty percent in trio designs and up to twenty-five percent in sibling designs, equivalent to doubling the effective sample size for trios. Simulations confirm the accuracy and unbiasedness of the calibrated estimator, and applications to UK Biobank family data demonstrate substantial precision gains and improved downstream Mendelian Randomization inference. Analysis of published within-sibship GWAS summary statistics further illustrates that the approach can be applied directly to publicly available data. Together, these results show that calibration provides a practical and powerful way to enhance family-based genetic analyses.
- Open Access
Effects of Distance Metrics and Scaling on the Perturbation Discrimination Score
The Perturbation Discrimination Score (PDS) is increasingly used to evaluate whether predicted perturbation effects remain distinguishable, including in Systema and the Virtual Cell Challenge. However, its behavior in high-dimensional gene-expression settings has not been examined in detail. We show that PDS is highly sensitive to the choice of similarity or distance measure and to the scale of predicted effects. Analysis of observed perturbation responses reveals that $\ell_1$ and $\ell_2$-based PDS behave very differently from cosine-based measures, even after norm matching. We provide geometric insight and discuss implications for future discrimination-based evaluation metrics.Open AccessEffects of Distance Metrics and Scaling on the Perturbation Discrimination Score
The Perturbation Discrimination Score (PDS) is increasingly used to evaluate whether predicted perturbation effects remain distinguishable, including in Systema and the Virtual Cell Challenge. However, its behavior in high-dimensional gene-expression settings has not been examined in detail. We show that PDS is highly sensitive to the choice of similarity or distance measure and to the scale of predicted effects. Analysis of observed perturbation responses reveals that $\ell_1$ and $\ell_2$-based PDS behave very differently from cosine-based measures, even after norm matching. We provide geometric insight and discuss implications for future discrimination-based evaluation metrics. - Open Access
- 28 July 2025
- Published by Springer Nature in Nature Communications
- Vol. 16  (1) , 1-10
Causal mediation analysis for time-varying heritable risk factors with Mendelian randomization
Understanding the causal mechanisms of diseases is crucial in clinical research. When randomized experiments are unavailable, Mendelian Randomization (MR) leverages genetic mutations to mitigate confounding. However, most MR analyses assume static risk factors, oversimplifying dynamic risk factor effects. The framework of life-course MR addresses this but struggles with limited GWAS cohort sizes and correlations across time points. We propose FLOW-MR, a computational approach estimating causal structural equations for temporally ordered traits using only GWAS summary statistics. FLOW-MR enables inference on direct, indirect, and path-wise causal effects, demonstrating superior efficiency and reliability, especially with noisy data. By incorporating a spike-and-slab prior, it mitigates challenges from extreme polygenicity and weak instruments. Applying FLOW-MR, we uncovered a childhood-specific protective effect of BMI on breast cancer and analyzed the evolving impacts of BMI, systolic blood pressure, and cholesterol on stroke risk, revealing their causal relationships.
- 28 July 2025
- Published by Springer Nature in Nature Communications
- Vol. 16  (1) , 1-10
Open AccessCausal mediation analysis for time-varying heritable risk factors with Mendelian randomization
Understanding the causal mechanisms of diseases is crucial in clinical research. When randomized experiments are unavailable, Mendelian Randomization (MR) leverages genetic mutations to mitigate confounding. However, most MR analyses assume static risk factors, oversimplifying dynamic risk factor effects. The framework of life-course MR addresses this but struggles with limited GWAS cohort sizes and correlations across time points. We propose FLOW-MR, a computational approach estimating causal structural equations for temporally ordered traits using only GWAS summary statistics. FLOW-MR enables inference on direct, indirect, and path-wise causal effects, demonstrating superior efficiency and reliability, especially with noisy data. By incorporating a spike-and-slab prior, it mitigates challenges from extreme polygenicity and weak instruments. Applying FLOW-MR, we uncovered a childhood-specific protective effect of BMI on breast cancer and analyzed the evolving impacts of BMI, systolic blood pressure, and cholesterol on stroke risk, revealing their causal relationships.
- 30 May 2025
- Published by Oxford University Press (OUP) in Biometrika
- Vol. 112  (4) , asaf038
Aggregating dependent signals with heavy-tailed combination tests
Summary: Combining dependent $ p $-values poses a long-standing challenge in statistical inference, particularly when aggregating findings from multiple methods to enhance signal detection. Recently, $ p $-value combination tests based on regularly-varying-tailed distributions, such as the Cauchy combination test and harmonic mean $ p $-value, have attracted attention for their robustness to unknown dependence. This paper provides a theoretical and empirical evaluation of these methods under an asymptotic regime where the number of $ p $-values is fixed and the global test significance level approaches zero. We examine two types of dependence among the $ p $-values. First, when $ p $-values are pairwise asymptotically independent, such as with bivariate normal test statistics with no perfect correlation, we prove that these combination tests are asymptotically valid. However, they become equivalent to the Bonferroni test as the significance level tends to zero for both one-sided and two-sided $ p $-values. Empirical investigations suggest that this equivalence can emerge at moderately small significance levels. Second, under pairwise quasi-asymptotic dependence, such as with bivariate $ t $-distributed test statistics, our simulations suggest that these combination tests can remain valid and exhibit notable power gains over the Bonferroni test, even as the significance level diminishes. These findings highlight the potential advantages of these combination tests in scenarios where $ p $-values exhibit substantial dependence. Our simulations also examine how test performance depends on the support and tail heaviness of the underlying distributions.- 30 May 2025
- Published by Oxford University Press (OUP) in Biometrika
- Vol. 112  (4) , asaf038
Aggregating dependent signals with heavy-tailed combination tests
Summary: Combining dependent $ p $-values poses a long-standing challenge in statistical inference, particularly when aggregating findings from multiple methods to enhance signal detection. Recently, $ p $-value combination tests based on regularly-varying-tailed distributions, such as the Cauchy combination test and harmonic mean $ p $-value, have attracted attention for their robustness to unknown dependence. This paper provides a theoretical and empirical evaluation of these methods under an asymptotic regime where the number of $ p $-values is fixed and the global test significance level approaches zero. We examine two types of dependence among the $ p $-values. First, when $ p $-values are pairwise asymptotically independent, such as with bivariate normal test statistics with no perfect correlation, we prove that these combination tests are asymptotically valid. However, they become equivalent to the Bonferroni test as the significance level tends to zero for both one-sided and two-sided $ p $-values. Empirical investigations suggest that this equivalence can emerge at moderately small significance levels. Second, under pairwise quasi-asymptotic dependence, such as with bivariate $ t $-distributed test statistics, our simulations suggest that these combination tests can remain valid and exhibit notable power gains over the Bonferroni test, even as the significance level diminishes. These findings highlight the potential advantages of these combination tests in scenarios where $ p $-values exhibit substantial dependence. Our simulations also examine how test performance depends on the support and tail heaviness of the underlying distributions.- Open Access
Validity and Interpretation of Two-Sample Mendelian Randomization with Binary Traits
Background: Two-sample Mendelian randomization (MR) is widely applied to binary exposures and outcomes. Yet standard MR models rely on linear effect assumptions that are difficult to interpret for binary traits. Although liability-based interpretations have been suggested, it remains unclear whether conventional summary-data MR is formally justified in this setting or what causal parameter it identifies. Methods: We develop a liability-threshold framework in which binary traits arise from underlying continuous liabilities. We derive explicit relationships between genome-wide association study (GWAS) coefficients obtained from logistic or linear regression on binary traits and marginal genetic associations on the liability scale. Under small genetic effects, typical for complex traits, observed-scale GWAS coefficients are approximately proportional to liability-scale associations. Results: This proportionality implies that standard two-sample MR methods remain statistically coherent for binary traits. MR applied to binary exposures or outcomes estimates a scaled causal effect between underlying liabilities rather than an effect on the observed binary scale. The scaling factor depends primarily on trait prevalence and is directly computable. Simulations and UK Biobank analyses confirm that, after rescaling, MR using binary traits recovers liability-scale causal effects consistent with analyses based on continuous traits. Conclusions: We provide a formal statistical justification for summary-data MR with binary traits and clarify the causal parameter being estimated. These results support routine MR practice for binary exposures and outcomes while enabling coherent interpretation of effect sizes. Key Messages: The interpretation of two-sample MR with binary exposures or outcomes is often unclear because GWAS analyses are performed on the observed binary scale. Under a liability threshold framework with small genetic effects, GWAS coefficients from logistic or linear regression on binary traits are approximately proportional to genetic associations on an underlying continuous liability scale. Consequently, conventional summary-data MR applied to binary or ordinal traits remains valid and estimates a scaled causal effect between liabilities, requiring no modification of existing methods.
Open AccessValidity and Interpretation of Two-Sample Mendelian Randomization with Binary Traits
Background: Two-sample Mendelian randomization (MR) is widely applied to binary exposures and outcomes. Yet standard MR models rely on linear effect assumptions that are difficult to interpret for binary traits. Although liability-based interpretations have been suggested, it remains unclear whether conventional summary-data MR is formally justified in this setting or what causal parameter it identifies. Methods: We develop a liability-threshold framework in which binary traits arise from underlying continuous liabilities. We derive explicit relationships between genome-wide association study (GWAS) coefficients obtained from logistic or linear regression on binary traits and marginal genetic associations on the liability scale. Under small genetic effects, typical for complex traits, observed-scale GWAS coefficients are approximately proportional to liability-scale associations. Results: This proportionality implies that standard two-sample MR methods remain statistically coherent for binary traits. MR applied to binary exposures or outcomes estimates a scaled causal effect between underlying liabilities rather than an effect on the observed binary scale. The scaling factor depends primarily on trait prevalence and is directly computable. Simulations and UK Biobank analyses confirm that, after rescaling, MR using binary traits recovers liability-scale causal effects consistent with analyses based on continuous traits. Conclusions: We provide a formal statistical justification for summary-data MR with binary traits and clarify the causal parameter being estimated. These results support routine MR practice for binary exposures and outcomes while enabling coherent interpretation of effect sizes. Key Messages: The interpretation of two-sample MR with binary exposures or outcomes is often unclear because GWAS analyses are performed on the observed binary scale. Under a liability threshold framework with small genetic effects, GWAS coefficients from logistic or linear regression on binary traits are approximately proportional to genetic associations on an underlying continuous liability scale. Consequently, conventional summary-data MR applied to binary or ordinal traits remains valid and estimates a scaled causal effect between liabilities, requiring no modification of existing methods.
- Open Access
Causal mediation analysis for time-varying heritable risk factors with Mendelian Randomization
Understanding the causal mechanisms of diseases is crucial in clinical research. When randomized experiments are unavailable, Mendelian Randomization (MR) leverages genetic mutations to mitigate confounding. However, most MR analyses assume static risk factors, oversimplifying dynamic risk factor effects. The framework of life-course MR addresses this but struggles with limited GWAS cohort sizes and correlations across time points. We propose FLOW-MR, a computational approach estimating causal structural equations for temporally ordered traits using only GWAS summary statistics. FLOW-MR enables inference on direct, indirect, and path-wise causal effects, demonstrating superior efficiency and reliability, especially with noisy data. By incorporating a spike-and-slab prior, it mitigates challenges from extreme polygenicity and weak instruments. Applying FLOW-MR, we uncovered a childhood-specific protective effect of BMI on breast cancer and analyzed the evolving impacts of BMI, systolic blood pressure, and cholesterol on stroke risk, revealing their causal relationships.
Open AccessCausal mediation analysis for time-varying heritable risk factors with Mendelian Randomization
Understanding the causal mechanisms of diseases is crucial in clinical research. When randomized experiments are unavailable, Mendelian Randomization (MR) leverages genetic mutations to mitigate confounding. However, most MR analyses assume static risk factors, oversimplifying dynamic risk factor effects. The framework of life-course MR addresses this but struggles with limited GWAS cohort sizes and correlations across time points. We propose FLOW-MR, a computational approach estimating causal structural equations for temporally ordered traits using only GWAS summary statistics. FLOW-MR enables inference on direct, indirect, and path-wise causal effects, demonstrating superior efficiency and reliability, especially with noisy data. By incorporating a spike-and-slab prior, it mitigates challenges from extreme polygenicity and weak instruments. Applying FLOW-MR, we uncovered a childhood-specific protective effect of BMI on breast cancer and analyzed the evolving impacts of BMI, systolic blood pressure, and cholesterol on stroke risk, revealing their causal relationships.
- Open Access
- 31 October 2023
- Published by Cold Spring Harbor Laboratory in Genome Research
- Vol. 33  (10) , 1774-1787
Infernape uncovers cell type–specific and spatially resolved alternative polyadenylation in the brain
Differential polyadenylation sites (PAs) critically regulate gene expression, but their cell type–specific usage and spatial distribution in the brain have not been systematically characterized. Here, we present Infernape, which infers and quantifies PA usage from single-cell and spatial transcriptomic data and show its application in the mouse brain. Infernape uncovers alternative intronic PAs and 3′-UTR lengthening during cortical neurogenesis. Progenitor–neuron comparisons in the excitatory and inhibitory neuron lineages show overlapping PA changes in embryonic brains, suggesting that the neural proliferation–differentiation axis plays a prominent role. In the adult mouse brain, we uncover cell type–specific PAs and visualize such events using spatial transcriptomic data. Over two dozen neurodevelopmental disorder–associated genes such asCsnk2a1andMecp2show differential PAs during brain development. This study presents Infernape to identify PAs from scRNA-seq and spatial data, and highlights the role of alternative PAs in neuronal gene regulation.
- 31 October 2023
- Published by Cold Spring Harbor Laboratory in Genome Research
- Vol. 33  (10) , 1774-1787
Open AccessInfernape uncovers cell type–specific and spatially resolved alternative polyadenylation in the brain
Differential polyadenylation sites (PAs) critically regulate gene expression, but their cell type–specific usage and spatial distribution in the brain have not been systematically characterized. Here, we present Infernape, which infers and quantifies PA usage from single-cell and spatial transcriptomic data and show its application in the mouse brain. Infernape uncovers alternative intronic PAs and 3′-UTR lengthening during cortical neurogenesis. Progenitor–neuron comparisons in the excitatory and inhibitory neuron lineages show overlapping PA changes in embryonic brains, suggesting that the neural proliferation–differentiation axis plays a prominent role. In the adult mouse brain, we uncover cell type–specific PAs and visualize such events using spatial transcriptomic data. Over two dozen neurodevelopmental disorder–associated genes such asCsnk2a1andMecp2show differential PAs during brain development. This study presents Infernape to identify PAs from scRNA-seq and spatial data, and highlights the role of alternative PAs in neuronal gene regulation.
- Open Access
Robust Statistical Inference for Cell Type Deconvolution
Cell type deconvolution is a computational method that estimates the proportions of different cell types within bulk transcriptomics data by leveraging information from reference single-cell RNA sequencing data. Despite its origin as a simple linear regression model, this approach faces challenges due to technical and biological variability and biases between the bulk and single-cell datasets. While several new methods have been developed, most only provide point estimates of cell type proportions, neglecting the uncertainty inherent in these estimates. Consequently, false positives can arise when comparing changes in cell type proportions across multiple individuals. In this paper, we introduce MEAD, a comprehensive statistical framework for efficient cell type deconvolution. Our approach constructs asymptotically valid confidence intervals for individual cell type proportions, as well as for quantifying changes in cell type proportions across multiple individuals. Our analysis accounts for factors such as biological variability in gene expressions, gene-gene dependence, cross-platform biases, and sequencing errors, without relying on parametric assumptions about the data distributions. Moreover, we establish necessary and sufficient conditions for identifying cell type proportions in the presence of platform-specific biases across sequencing technologies.
Open AccessRobust Statistical Inference for Cell Type Deconvolution
Cell type deconvolution is a computational method that estimates the proportions of different cell types within bulk transcriptomics data by leveraging information from reference single-cell RNA sequencing data. Despite its origin as a simple linear regression model, this approach faces challenges due to technical and biological variability and biases between the bulk and single-cell datasets. While several new methods have been developed, most only provide point estimates of cell type proportions, neglecting the uncertainty inherent in these estimates. Consequently, false positives can arise when comparing changes in cell type proportions across multiple individuals. In this paper, we introduce MEAD, a comprehensive statistical framework for efficient cell type deconvolution. Our approach constructs asymptotically valid confidence intervals for individual cell type proportions, as well as for quantifying changes in cell type proportions across multiple individuals. Our analysis accounts for factors such as biological variability in gene expressions, gene-gene dependence, cross-platform biases, and sequencing errors, without relying on parametric assumptions about the data distributions. Moreover, we establish necessary and sufficient conditions for identifying cell type proportions in the presence of platform-specific biases across sequencing technologies.