About

Victor Wenze Zhong is associated with the Bronfenbrenner Center for Translational Research at Cornell University. The center assists faculty in developing translational research projects, providing support such as proposal preparation, training, technical assistance, and fostering collaborative relationships. The center also offers workshops, summer institutes, and talks on current research topics, aiming to facilitate the dissemination and funding of translational research efforts. Specific details about Dr. Zhong's individual research focus, background, or key contributions are not provided on the page.

Research topics

• Internal medicine
• Medicine
• Environmental health
• Food science
• Pathology
• Demography
• Gerontology
• Endocrinology

Selected publications

• Phenome‐ and Genome‐Wide Insights Into Dietary Patterns: Health Consequences and Genetic Basis

  eFood · 2026-05-13

  articleOpen access

  ABSTRACT Dietary patterns (DPs) integrate complex eating behaviors and provide stronger insights into health than analyses of single nutrients. However, most studies have examined only a limited range of outcomes, and the biological mechanisms underlying DPs remain poorly defined. Using a robust data‐driven approach, we derived three major DPs from dietary recall data and evaluated their associations with > 1000 diseases and biomarkers. Mediation analyses investigated metabolic intermediates, and genome‐wide association studies (GWAS) identified genetic loci shaping DPs. DP1 represented a plant‐forward, nutrient‐dense profile; DP2 was meat‐and‐vegetables oriented; and DP3 was characterized by energy‐dense snack intake. In phenome‐wide association analyses, DP1 was broadly protective, inversely associated with over 50 diseases, whereas DP3 was detrimental, linked to increased risk across more than 60 outcomes. DP2 showed limited associations. Mediation analyses indicated that the benefits of DP1 were partly attributable to higher circulating polyunsaturated fatty acids, while the harms of DP3 were mediated through lower high‐density lipoprotein levels. GWAS revealed distinct genetic underpinnings, linking DP1 to APOE , APOC1 , and NEGR1 , and DP3 to the MAPT locus. Together, these findings provide a comprehensive characterization of DPs, reveal metabolic and genetic pathways underlying diet‐disease relationships, and highlight opportunities for advancing personalized nutrition and disease prevention.

  PublisherDOI

• Association of eating duration less than 8 h with all-cause, cardiovascular, and cancer mortality

  Diabetes & Metabolic Syndrome Clinical Research & Reviews · 2025-07-01 · 5 citations

  articleSenior authorCorresponding
  PublisherDOI

• Biological age acceleration and interaction with genetic predisposition in the risk of type 2 diabetes and coronary artery disease

  GeroScience · 2025-04-29 · 4 citations

  articleOpen accessSenior author
  PublisherOA PDFDOI

• Abstract P3029: Phenotypic and Genetic Correlations of Imaging Traits in the Abdomen-Heart-Brain Axis: Insights into Multimorbidity of Cardiometabolic Diseases and Depression

  Circulation · 2025-03-11

  articleSenior author

  Objective: The development of multimorbidity of cardiometabolic diseases (CMDs) and depression likely involves pathophysiological processes across multiple organs. This study aimed to investigate imaging traits in the abdomen-heart-brain axis linked to the CMDs-depression multimorbidity and assess their phenotypic and genetic connections using magnetic resonance imaging (MRI) data. Methods: This study analyzed multi-organ MRI data including 7 abdominal, 82 cardiac, and 676 brain traits from 31,839 UK Biobank participants. CMDs included coronary artery disease, stroke, heart failure, type 2 diabetes, hypertension and hyperlipidemia. First, multivariable logistic regressions estimated associations of imaging traits with the CMDs-depression multimorbidity. Second, multivariable linear regressions quantified pairwise correlations among significant cross-organ traits. Third, transcriptome-wide association study (TWAS) identified shared genes of traits significantly correlated with each other across all 3 regions (i.e., triads). Fourth, LASSO regressions constructed genomics scores using single nucleotide polymorphisms from the shared genes. Fifth, the prediction performance of imaging traits and genomics scores for the CMDs-depression multimorbidity was evaluated by AUCs. FDR correction was used to adjust for multiple testing. Results: Seven abdominal, 27 cardiac, and 340 brain traits were significantly associated with the CMDs-depression multimorbidity. Among 11,749 pairs of significant cross-organ imaging traits, 6385 abdomen-heart-brain triads were identified, with liver volume as the most connected node. The directions of associations among imaging traits in these cross-organ triads aligned with directions of their biological functions, suggesting multiple organs acting as a coordinated system. TWAS revealed TNFSF12 from whole blood as the shared gene underlying the significant liver-heart-brain connection associated with the CMDs-depression multimorbidity. The integration of imaging traits with genomics score (AUC = 0.84) improved prediction performance for the CMDs-depression multimorbidity, outperforming demographic and lifestyle models (AUC = 0.78) (P = 0.014). Conclusions: This study identified highly correlated cross-organ imaging traits that were associated with the CMDs-depression multimorbidity. A shared gene was found for a liver-heart-brain triad. Imaging traits and genomic score significantly enhanced prediction performance on top of traditional factors.

  PublisherDOI

• Met246 and Asn250 in the D2 protein are essential for the operation of the quinone-Fe-acceptor complex of Photosystem II

  Plant and Cell Physiology · 2025-07-14 · 2 citations

  articleOpen access1st authorCorresponding

  The chemical properties of the primary (QA) and secondary (QB) plastoquinone electron acceptors of Photosystem II (PS II) depend on their protein environments. The DE loop of the D2 protein (residues 222-262) contributes to the QA-binding site while the DE loop of the D1 protein (residues 233-266) contributes to the QB-binding environment. The roles of the invariant D2-Met246 and D2-Asn250 residues in the vicinity of the QA-binding site have been investigated in the cyanobacterium Synechocystis sp. PCC 6803 using mutants targeting both residues. The M246F strain was phenotypically similar to control cells; however, the M246A, N250A, and N250H strains had slowed photoautotrophic growth and were sensitive to high light and the addition of formate. In addition, the M246K and N250N strains were unable to assemble PS II. Chlorophyll a fluorescence measurements indicated electron transfer between QA and QB was modified in the M246A, N250A, and N250H strains, and the exchange of plastoquinol between the QB-binding site and the plastoquinone pool in the thylakoid membrane was impaired. Modified electron transfer in these mutants in the presence or absence of formate was restored by the addition of bicarbonate. In addition, thermoluminescence measurements showed a down shift in the redox midpoint potential of the QA/QA- couple in the N250A and N250H strains. These results demonstrate that Met246 and Asn250 play indispensable roles in the quinone-iron-acceptor complex, influencing both QA binding and the binding of the bicarbonate ligand to the non-heme iron that is located between QA and QB.

  PublisherOA PDFDOI

• Association of blood pressure levels in the nonhypertensive range with mortality among people without traditional risk factors

  Journal of Hypertension · 2025-12-17

  articleSenior author

  BACKGROUND: While emerging evidence suggests guideline-defined nonhypertensive blood pressure (BP) may encompass heterogeneous risk, the relationship between BP variations within nonhypertensive ranges and mortality risk remains inadequately characterized among individuals without traditional cardiovascular risk factors. This study investigated whether nonhypertensive range of SBP, DBP, and pulse pressure (PP) are associated with long-term mortality in a healthy population. METHODS: This study included 80 730 UK Biobank participants without traditional cardiovascular risk factors and with nonhypertensive BP (SBP <140 mmHg, DBP <90 mmHg, and PP <60 mmHg). Participants were followed up for all-cause, cardiovascular, and noncardiovascular mortality. Associations were assessed using multivariable Cox proportional hazards models with restricted cubic splines. RESULTS: Over a median follow-up of 13.7 years, 2553 deaths occurred. SBP and PP showed significant nonlinear associations with all-cause mortality ( P -overall <0.01), while DBP showed a linear inverse association ( P -overall = 0.049). Compared to the third quintile, the lowest PP quintile (<40 mmHg) was associated with 26% higher mortality risk (hazard ratio 1.26, 95% confidence interval [95% CI] 1.10-1.44), and the highest quintile (53-60 mmHg) with 14% higher risk (hazard ratio 1.14, 95% CI 1.01-1.28). The lowest SBP quintile (<114 mmHg) was associated with 16% higher risk (hazard ratio 1.16, 95% CI 1.02-1.32) compared to the third quintile (120-126 mmHg). CONCLUSION: Even within nonhypertensive ranges, the lowest and highest quintiles of PP level, as well as low-normal SBP and DBP levels, were associated with increased mortality risk in a healthy population.

  PublisherDOI

• ModelTables: A Corpus of Tables about Models

  ArXiv.org · 2025-12-18

  preprintOpen access

  We present ModelTables, a benchmark of tables in Model Lakes that captures the structured semantics of performance and configuration tables often overlooked by text only retrieval. The corpus is built from Hugging Face model cards, GitHub READMEs, and referenced papers, linking each table to its surrounding model and publication context. Compared with open data lake tables, model tables are smaller yet exhibit denser inter table relationships, reflecting tightly coupled model and benchmark evolution. The current release covers over 60K models and 90K tables. To evaluate model and table relatedness, we construct a multi source ground truth using three complementary signals: (1) paper citation links, (2) explicit model card links and inheritance, and (3) shared training datasets. We present one extensive empirical use case for the benchmark which is table search. We compare canonical Data Lake search operators (unionable, joinable, keyword) and Information Retrieval baselines (dense, sparse, hybrid retrieval) on this benchmark. Union based semantic table retrieval attains 54.8 % P@1 overall (54.6 % on citation, 31.3 % on inheritance, 30.6 % on shared dataset signals); table based dense retrieval reaches 66.5 % P@1, and metadata hybrid retrieval achieves 54.1 %. This evaluation indicates clear room for developing better table search methods. By releasing ModelTables and its creation protocol, we provide the first large scale benchmark of structured data describing AI model. Our use case of table discovery in Model Lakes, provides intuition and evidence for developing more accurate semantic retrieval, structured comparison, and principled organization of structured model knowledge. Source code, data, and other artifacts have been made available at https://github.com/RJMillerLab/ModelTables.

  PublisherOA PDFDOI

• Epigenetic Signatures of Type 2 Diabetes Remission and Substantial Weight Loss: The Role of DNA Methylation in Response to Low-Calorie Diet Intervention

  Journal of the American Nutrition Association · 2025-06-05

  articleSenior authorCorresponding

  OBJECTIVES: Low-calorie diet (LCD) interventions can lead to remission of type 2 diabetes (T2D), but the underlying mechanisms are not well understood. As a diet-sensitive regulator of gene expression, DNA methylation may reveal pathways underlying remission. However, whether individuals with different responses to LCD-induced T2D remission and weight loss exhibit distinct DNA methylation patterns remains unclear. METHODS: A 3-month intensive weight loss intervention (815-835 kcal/d) and a following 3-month weight loss maintenance phase were conducted among participants with T2D. DNA methylation was measured using the Infinium Methylation EPIC BeadChip (935K) at 4 timepoints. Differentially methylated cytosine-phosphate-guanine (CpG) sites and regions were analyzed to identify changes between those who achieved T2D remission and weight loss >12 kg and those who did not. Predictive models based on DNA methylation profiles at baseline and week 1 were developed to forecast individual responses to LCD. RESULTS: . Pathway analysis mapped these genes to glucose and lipid metabolism pathways. Methylation-inferred scores showed a greater improvement in predicting T2D remission and weight loss status following the LCD intervention than the base models (area under curve ranges: 0.86-0.88 vs. 0.73-0.80). CONCLUSIONS: Variation in DNA methylation profiles across individuals with differing responses to the LCD, highlighting the epigenetic roles in the effects of the LCD on weight loss and T2D remission. Baseline DNA methylation status may serve as a predictor of T2D remission in response to LCD intervention. CLINICAL TRIAL REGISTRY NUMBER AND WEBSITE: This trial was registered at the https://clinicaltrials.gov/study/NCT05472272?cond=NCT05472272&rank=1 as NCT05472272.

  PublisherDOI

• Interaction between dietary cholesterol intake and genetic variants for cholesterol absorption in relation to coronary artery disease mortality: a prospective analysis in the Singapore Chinese Health Study

  American Journal of Clinical Nutrition · 2025-09-09

  articleCorresponding
  PublisherDOI

• Abstract 062: Disease Trajectories of Cardiometabolic Diseases and Depression: Transition Patterns, Multiomics Signatures, Prognosis and Prediction

  Circulation · 2025-03-11

  articleSenior author

  Objective: Cardiometabolic diseases (CMDs) and depression, among the most prevalent physical and mental diseases, frequently co-occur and are associated with a higher risk of premature mortality. However, the trajectories of their multimorbidity occurrence, underlying biological mechanisms, and early prediction remain poorly understood. Methods: This study was conducted on 467,592 UK Biobank participants without baseline CMDs and depression. CMDs included type 2 diabetes, coronary artery disease, stroke, and heart failure. Multistate models were used to investigate transition probabilities and identify multiomics signatures for transitions from baseline to single morbidity and to multimorbidity. Multiomics prediction models were constructed using least absolute shrinkage and selection operator Cox regression. Model performance was evaluated by the change of Harrell's C statistic (Δ C-statistic) compared to the demographic model and area under receiver-operating characteristic curves (AUCs). Results: During a median follow-up of 14.6 years, 64,442 participants developed CMDs alone, 17,533 developed depression alone, and 6104 developed multimorbidity. Depression preceding CMDs showed a 2.5% higher 15-year multimorbidity probability and 1.2% higher 5-year mortality risk than the reverse sequence. Multimorbidity was associated with a 14-26% higher mortality risk and 3.6-3.8 years shorter survival time compared to those without CMDs or depression. Distinct and shared multiomics signatures underlying disease trajectories were identified. Valine, leucine, and isoleucine biosynthesis and cytokine-cytokine receptor interaction pathways were implicated in both CMDs and depression progression. Proteomics scores showed superior prediction performance across nine disease transitions, with Δ C-statistic ranging from 0.07 for health-death to 0.22 for health-depression-CMDs-death, compared to the genomics (0-0.04) and metabolomics (0-0.16) scores. For 15-year outcome prediction, the proteomics scores model achieved AUCs of 0.65-0.87 for various transitions, significantly outperforming traditional risk factors-based and other omics models. Conclusions This study revealed distinct transition patterns, identified associated multiomics signatures, and constructed multiomics prediction models for the multimorbidity cluster of CMDs and depression.

  PublisherDOI

Frequent coauthors

• Elizabeth J. Mayer‐Davis

  26 shared

• Caiming Xiong

  19 shared

• Donald M. Lloyd‐Jones

  Northwestern University

  17 shared

• Luke Zettlemoyer

  16 shared

• Norrina B. Allen

  Northwestern University

  16 shared

• Hongyan Ning

  Northwestern University

  16 shared

• Richard Socher

  16 shared

• John T. Wilkins

  Northwestern University

  15 shared