Yue-Qing Hu

Zero inflation remains a major challenge in microbiome differential abundance analysis, often resulting in inflated type I error rates or reduced statistical power. Although numerous methods have been proposed to address excess zeros, many existing approaches do not explicitly distinguish between distinct zero-generating mechanisms, and censoring-based modeling perspectives remain relatively underexplored in microbiome data analysis. To advance methodological development in this area, we introduce a novel censoring-based modeling perspective for microbiome differential abundance analysis. Specifically, we develop a joint mixture Tobit (joint mTobit) method for zero-inflated microbiome data. Building on a mixture Tobit formulation, the model incorporates a point-mass component to represent structural zeros, while modeling latent microbial abundance via a Tobit regression component. The proposed framework jointly links disease status, latent true microbial abundance, and relevant covariates within a unified probabilistic model, allowing appropriate adjustment for confounding factors and facilitating more reliable statistical inference. By explicitly modeling the latent abundance underlying observed counts, the joint mTobit method improves estimation stability and enhances detection power under zero inflation. Extensive simulation studies demonstrate that the joint mTobit method achieves effective type I error control while maintaining high statistical power and stable coefficient estimation across a wide range of settings. Application to a real-world colorectal cancer and adenoma microbiome dataset further illustrates its ability to identify biologically meaningful differentially abundant taxa. Overall, this work develops a joint mTobit modeling framework for zero-inflated microbiome data, enabling inference on latent microbial abundance and providing a censoring-based statistical framework for investigating disease–microbiome associations in differential abundance analysis.

Motivation: Mendelian randomization (MR) is widely used for causal inference using genetic data, yet most existing MR methods treat exposures as static and have limited capacity to analyze longitudinal measurements. In many biomedical studies, multiple exposures evolve over time, are strongly correlated, and share unobserved temporal structure. Ignoring these features can lead to unstable estimation and reduced power. Results: We propose factor-augmented multivariable functional Mendelian randomization (FA-MFMR) for estimating time-varying causal effects of multiple longitudinal exposures. FA-MFMR integrates genetic instruments with functional representations of exposure trajectories and a low-dimensional factor structure to capture shared temporal patterns. By separating common temporal variation from exposure-specific components, it mitigates multicollinearity, improves estimation stability, and accommodates sparse causal structures in which some exposures have no effect throughout the observation period. Simulations demonstrate lower estimation error and more efficient uncertainty quantification than competing approaches, particularly for highly correlated and sparsely observed longitudinal exposures. Applied to longitudinal biomarkers in a Parkinson’s disease cohort, FA-MFMR characterizes age-dependent patterns in the estimated trajectory-level coefficient functions while highlighting uncertainty in localized effects. Availability and implementation: The developed R package and code to reproduce all the results are available at https://github.com/YQHuFD/FA-MFMR.

Multitrophic interactions can strongly influence the structure and functioning of ecosystems, but how plant diversity influences the direction and predictability of multitrophic interactions across agricultural and natural ecosystems remains unclear. Using 149 field studies across five continents, we found that, on average, increasing plant diversity tended to exert differential top-down and bottom-up effects in croplands versus grasslands and forests. Organic and nonorganic croplands exhibited 846 and 148% higher invertebrate natural enemy-to-herbivore abundance ratios under increased plant diversity, consistent with top-down control patterns where predator gains cause herbivore declines, enhancing crop outcomes. In grasslands and forests, increasing plant diversity was associated with bottom-up effects where enhanced productivity increased both herbivore and predator populations, with the enemy-to-herbivore ratio increasing 4.73% for grasslands and 21.2% for forests. Our findings suggest that biodiversity effects on productivity are not solely explained by direct plant-plant interactions and the resulting biodiversity-productivity relationship. Rather, they reveal patterns consistent with the framework of top-down and bottom-up effects, the relative balance of which may vary depending on ecosystem and management type. The magnitude of the effects of diversified farming on crop pests suggests that crop diversification may be an important avenue for managing crop pests preventatively and thereby enhancing agricultural sustainability.

Inferring causal effects with unmeasured confounder is a main challenge in causal inference. Many researchers impose parametric assumptions on the distribution of unmeasured confounder. However, due to the unobservable nature of the unmeasured confounder, it is more reasonable to leave its distribution unrestricted. Another key challenge in causal inference is the involvement of invalid instrumental variables, which may lead to biased inference and potentially deceptive results. To this end, we employ a flexible semiparametric model that allows for possibly invalid instruments without specifying the distribution of unmeasured confounder in this work. A penalized semiparametric estimator for causal effects is constructed and its oracle and asymptotic properties are well established for statistical inference. We evaluate the performance of the estimator through simulation studies, revealing that our proposed estimator exhibits asymptotic unbiasedness and robustness in estimating causal effects, along with consistent selection of invalid instruments. We also demonstrate its application using Atherosclerosis Risk in Communities Study data set, which further validates its robustness in the presence of invalid instruments.

The diversity–productivity relationship suggests that increasing plant species could increase primary productivity, with this effect being explained in part by the suppression of plant antagonists. We conducted a global synthesis of 609 studies to investigate how plant diversity affects plants and their antagonists. Here we show that increasing plant species consistently promotes plant performance and suppresses antagonist performance in agro-ecosystems, grasslands and forests, for herbaceous and woody plants, across tropical and temperate zones, and for replacement series and additive experimental design studies. Crop diversification (for example, intercropping and cover cropping) indirectly promotes crop production through the suppression of pests. This shows that diversifying planting systems can increase productivity while reducing reliance on synthetic pesticides, offering a sustainable pathway for agriculture from subsistence to large-scale agriculture. Overall, these results suggest that crop diversification has considerable potential to support sustainable agro-ecosystems that benefit productivity while reducing reliance on synthetic pesticides.

Gestational diabetes mellitus (GDM) has a high heritability and frequently co-occurs with type 2 diabetes (T2D), indicating shared genetic mechanisms. Firstly, we employed RHOGE to ascertain the genetic correlation and putative causal directions between GDM and T2D. Subsequently, the Genotype-Tissue Expression Project v8 eQTls files and the FinnGen R11 dataset were employed to conduct cross-tissue transcriptome association studies, Functional Summary-based Imputation in single tissues, and Gene Analysis combined with Multimarker Analysis of Genomic Annotation for GDM and T2D, respectively. A total of 5 genes were identified as GDM susceptibility and 97 genes linked to T2D susceptibility. Of these, four genes (COBLL1, NRBP1, IFT172 and TRIM54) were identified as being shared. Mendelian randomization and colocalization analyses revealed the causal associations of them with GDM and T2D in distinct tissues. Subsequent analyses indicated COBLL1 may influence GDM and T2D risk by regulation of actin filament polymerization and interactions with chemical responses. NRBP1 may confer protective effects against diabetes through regulation of insulin secretion, while IFT172 and TRIM54 may play a role in energy balance signaling and metabolism of skeletal muscle, respectively. Our study provides insight into the shared genetic mechanism between GDM and T2D and identifies potential targets for pharmacological intervention.

Causal inference is an essential approach for understanding biological processes. Traditional causal inference methods assume a linear relationship between different biological traits, whereas their true causal relationship may be nonlinear, such as U-shaped. Moreover, when the instrument set includes weak and pleiotropic genetic instruments, accurately capturing the shape of these relationships becomes challenging. To address these issues, we propose model-averaged control function-based instrumental variable regression, a two-stage framework based on a model-averaged control function approach to estimate the marginal effect function, which represents the derivative of the causal relationship. In the first stage, a model averaging technique is employed to estimate the control function, thereby reducing weak genetic instrument bias. In the second stage, B-spline approximation is applied to estimate the marginal effect function, while SCAD penalization is used to minimize pleiotropic instrument bias. We establish the asymptotic properties of the proposed estimator and demonstrate its robust performance through simulations. Application to the Atherosclerosis Risk in Communities dataset highlights a nonlinear causal relationship between body mass index and hypertension, with the proposed method effectively estimating the specific shape and trend of the relationship.

Background: Coronary heart disease (CHD) is the leading cause of premature mortality. The incremental value of a polygenic risk score (PRS) to a clinical risk score towards improving CHD prediction is controversial. Meanwhile, the effect of PRSs on CHD prediction in the chronic disease population is unclear. Methods: Utilizing publicly available summary statistical data, we developed several PRSs using the genome data of European ancestry from the Atherosclerosis Risk in Communities Study. Furthermore, we investigated the association of CHD with the best-performing PRS in both the overall and chronic disease cohorts. Additionally, we evaluated whether adding the best-performing PRS to the clinical risk score improves risk prediction. Results: A total of 6152 subjects (767 CHD cases) were included in this study. The high values from the developed best-performing PRS were significantly associated with an increased risk of CHD, with a stronger association in the hypertensive population (interaction p = 0.0144). Compared with individuals in the bottom 20% of the PRS values, those in the top 20% were more than 3-fold more likely to develop CHD in the overall cohort, rising to 5-fold in the hypertensive cohort. Adding PRS to the clinical risk score significantly improved the C-index (0.72 to 0.74; p = 0.004), with a 10% net reclassification improvement overall. The hypertensive population showed the greatest improvements. Furthermore, we observed a significant gradient of 10-year and lifetime risk of CHD based on the PRS within each clinical risk category. Conclusions: Compared to the clinical risk score, integrating the PRS significantly improved CHD prediction and better identified CHD risk trajectories, especially in the European hypertensive adult population.

Background: Coronary heart disease (CHD) is the leading cause of premature mortality. The incremental value of polygenic risk score (PRS) to clinical risk score towards improving CHD prediction is controversial. Meanwhile, the effect of PRS on CHD prediction in the chronic disease population is unclear. Methods: Utilizing publicly available summary statistic data, we developed several PRSs using genome data of European ancestry from the Atherosclerosis Risk in Communities Study. Furthermore, we investigated the association of CHD with best-performing PRS in both the overall and chronic disease cohorts. Additionally, we evaluated whether adding best-performing PRS to clinical risk score improves the risk prediction. Results: A total of 6,152 subjects (767 CHD cases) were included in this study. The developed best-performing PRS was significantly associated with increased risk of CHD, with a stronger association in the hypertensive population (interaction P = 0.0144). Compared with individuals in the bottom 20% of PRS, those in the top 20% were more than 3-fold more likely to develop CHD in overall cohort, rising to 5-fold in the hypertensive cohort. Adding PRS to the clinical risk score significantly improved the C-index (0.72 to 0.74; P = 0.004) with a 10% net reclassification improvement overall. The hypertensive population showed the greatest improvements. Furthermore, we observed significant gradient of 10-year and lifetime risk of CHD based on PRS within each clinical risk category. Conclusions: Compared to the clinical risk score, integrating PRS significantly improved CHD prediction and better identified CHD risk trajectories, especially in the hypertensive population.

Longitudinal human microbial data offer insights into microbiome dynamics over time. Traditional methods usually overlook temporal relationships among samples from the same subject. Here, we presented the Longitudinal Microbial Data Distance (LorDist) method, which uses functional data fitting to construct a distance matrix integrating information from the same subject at different time points. Simulation data showed that LorDist handled well up to 60% sparseness and worked robustly with various sequencing depths and time points. Empirical data analysis demonstrated that LorDist excels in capturing differences across subjects with longitudinal microbiome data. LorDist presented the potential of longitudinal microbial data in addressing temporal autocorrelation and distinguishing phenotypes. IMPORTANCE: Longitudinal analysis of the human microbiome is critical for understanding its dynamic role in health and disease. However, current analytical approaches struggle to address key challenges, such as data sparsity and irregular sampling, inherent to time-series microbiome studies. Here, we developed longitudinal microbial data distance (LorDist), an innovative method leveraging functional data analysis to model temporal microbial dynamics with enhanced precision. Compared to existing methods, LorDist consistently outperforms in discerning biologically meaningful group differences, even in highly sparse data sets or under fluctuating sequencing depths. Our findings demonstrate LorDist’s robust performance on real-world data sets involving inflammatory bowel disease and infant gut development. By explicitly preserving the temporal structure inherent in microbiome data, LorDist enables robust detection of subtle yet critical biological shifts, paving the way for improved diagnostics and personalized therapeutic strategies in microbiome science.