JYJia You

Jia You

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

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


Jia You has a total of 161 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Self-Evolving Agent for Longitudinal Personal Health Management

    by Hongcheng Guo, Haoran Li, Jiebi Deng, Tong Jin, Jinghong Han, Yuxin Wang, Zexin Wang, Jinyu Sun, Qingyi Si, Weikang Gong, et al.

    Medical large language models and agents show strong single-encounter question answering and evidence interpretation but struggle to accumulate and govern experience across encounters. We present HealthJev, an open-source agent separating shared safety rules and medical knowledge from a private longitudinal memory and distilling lessons from completed cases. To filter out erroneous lessons before they affect future behaviour, we self-train JevGate, a lightweight System-One model that decides which lessons enter memory; each decision takes a median of 86 milliseconds and about 1,000 input tokens, sustaining experience accumulation at negligible run-time overhead. On a longitudinal benchmark of 90-day simulated daily interactions for 20 users built from real MIMIC-IV trajectories, HealthJev outperformed a memory-free ReAct agent and three memory-augmented frameworks, with all paired comparisons significant. Across six tasks spanning clinical records, medical imaging, genomics and clinical text, HealthJev improved every task with a mean gain of 13.3 percentage points over the base condition, surpassing all baselines. The frozen library transferred to six further backbones with macro-average gains of 6.8 to 15.1 percentage points, and one task's library brought gains on other tasks. These results indicate that governed, self-evolving memory supports longitudinal personal health management beyond single encounters. HealthJev is publicly available at https://github.com/HC-Guo/Self-Evolving_HealthJev.

  • Open Access

    OmicFormer: a statistical priors–informed transformer for accurate and generalizable omics prediction of diseases and complex traits

    by Weikang Gong, Hanlin Jiang, Chang Yang, Menghan Qin, Jia You, Jianfeng Feng, Jin-Tai Yu, Wei Cheng

    Precision medicine faces a critical challenge in translating high-dimensional omics data into robust disease predictions across diverse populations. Current approaches often fail under distribution shifts, partly due to their inability to encode complex biological feature dependencies. We present OmicFormer, a Transformer-based architecture that embeds two complementary statistical priors, i.e., feature–label associations and feature–feature dependencies, directly into its representation learning. This design captures local and long-range omic interactions often missed by conventional methods. Analyzing 500,000 UK Biobank participants, OmicFormer significantly outperforms strong baselines across 450 disease and 900 trait prediction tasks, with substantial gains spanning diverse metabolic, neurological, cardiovascular, and gastrointestinal conditions, alongside enhanced prediction of circulating metabolites, bone density traits, and retinal imaging biomarkers. Crucially, OmicFormer demonstrates robust generalization, achieving a substantial improvement over tree-based methods in an independent proteomics cohort across 19 diseases (GNPC, N = 7,289), and outperforming tree-based models across 50 multi-site neuroimaging sites (N = 4,728) for autism and schizophrenia classification. By explicitly embedding statistical structure, OmicFormer provides an interpretable and generalizable foundation for omics-based precision medicine.

  • Open Access

    Analysis of whole genome sequencing and plasma metabolomics unveil genetic determinants and clinical implications for human health

    by Yi-Xuan Wang, Yi-Xuan Qiang, Yi-Jun Ge, Yue-Ting Deng, Bang-Sheng Wu, Liu Yang, Yi-Lin Chen, Xiao-Yu He, Yu He, Bing-Ran Yao, et al.

    Blood metabolites are critical biomarkers linking genetic variation to diverse health outcomes, exhibiting significant variability across individuals. Here, we analyze and integrate whole-genome sequencing with plasma metabolomics to investigate the genetic architecture of 313 metabolic biomarkers in up to 199,138 UK Biobank participants. Through single-variant analyses, we identify 36,105 independent signals, with 22.20% being novel. Furthermore, we pinpoint 12,361 putative causal variant-trait associations, demonstrating enhanced causal signal discovery and improved fine-mapping resolution compared with imputed array-based approaches. Rare-variant aggregate testing reveals 1,527 conditionally independent protein-coding gene-trait pairs, with 32.9% being unique to non-coding regions. We estimate the heritability at a median of h2 = 0.31, nearly tripling the heritability estimates obtained from array-based approaches. Integrating our findings with disease genetics reveals 245 potential causal associations, such as that between omega-3 fatty acids and cholelithiasis risk. Prioritizing proteins with concordant effects on metabolites and clinical outcomes revealed 410 potential targets, offering opportunities for therapeutic strategies and drug repurposing. Our open-access resource (https://metabolome-whole-genome-landscape.com/) provides a foundation for future research into metabolic pathways, disease mechanisms, and therapeutic development.


  • Plasma proteomic profiles linked to suicidal behaviors

    by Bei Zhang, Jia You, Edmund T. Rolls, Peng Ren, Yuzhu Li, Wei Zhang, Barbara J. Sahakian, Fei Li, Jianfeng Feng, Wei Cheng


  • Open Access

    Genetic architecture of plasma metabolome in 254,825 individuals

    by Yi-Xuan Qiang, Yi-Xuan Wang, Xiao-Yu He, Yue-Ting Deng, Yi-Jun Ge, Bang-Sheng Wu, You Jia, Jian-Feng Feng, Wei Cheng, Jin-Tai Yu

    Circulating metabolites are crucial to biological processes underlying health and diseases, yet their genetic determinants remain incompletely understood. Here, we investigate the genetic architecture of nuclear magnetic resonance-based metabolomics, analyzing 249 metabolic measures and 64 biologically plausible ratios in 254,825 participants. We conduct a genome-wide association study (GWAS) identifying 24,438 independent variant-metabolite associations across 427 loci, with effect sizes highly concordant with 19 previous studies. Fine-mapping pinpoints 3610 putative causal associations, 785 of which are novel. Additionally, we utilize whole exome sequencing data and uncover 2948 gene-metabolite associations through aggregate testing, underscoring the importance of rare coding variants overlooked in GWAS. Integrating our findings with disease genetics reveals potential causal associations, such as between acetate levels and the risk of atrial fibrillation and flutter. Collectively, this study delineates the complex genetic architecture of the plasma metabolome, offering a valuable resource for future investigations into disease mechanisms and therapeutic strategies.


  • Top Cited Papers
    Open Access

    Mapping the plasma metabolome to human health and disease in 274,241 adults

    by Jia You, Xi-Han Cui, Yi-Lin Chen, Yi-Xuan Wang, Hai-Yun Li, Yi-Xuan Qiang, Ji-Yun Cheng, Yue-Ting Deng, Yu Guo, Peng Ren, et al.

    A systematic characterization of metabolic profiles in human health and disease enhances precision medicine. Here we present a comprehensive human metabolome–phenome atlas, using data from 274,241 UK Biobank participants with nuclear magnetic resonance metabolic measures. This atlas links 313 plasma metabolites to 1,386 diseases and 3,142 traits, with participants being prospectively followed for a median of 14.9 years. This atlas uncovered 52,836 metabolite–disease and 73,639 metabolite–trait associations, where the ratio of cholesterol to total lipids in large low-density lipoprotein percentage was found as the metabolite associated with the highest number ( n = 526) of diseases. In addition, we found that more than half (57.5%) of metabolites showed statistical variations from healthy individuals over a decade before disease onset. Combined with demographics, the machine-learning-based metabolic risk score signified the top 30 (around 10%) metabolites as biomarkers, yielding favourable classification performance (area under the curve > 0.8) for 94 prevalent and 81 incident diseases. Finally, Mendelian randomization analyses provided support for causal relationships of 454 metabolite–disease pairs, among which 402 exhibited shared genetic determinants. Additional insights can be gleaned via an accessible interactive resource ( https://metabolome-phenome-atlas.com/ ).


  • Machine learning-assisted optimization of dietary intervention against dementia risk

    by Si-Jia Chen, Hui Chen, Jia You, Shi-Dong Chen, Yan Fu, Wei Zhang, Liyan Huang, Jian-Feng Feng, Xiang Gao, Wei Cheng, et al.

    A healthy diet has been associated with a reduced risk of dementia. Here we devised a Machine learning-assisted Optimizing Dietary intERvention against demeNtia risk (MODERN) diet based on data from 185,012 UK Biobank participants, 1,987 of whom developed all-cause dementia over 10 years. We first identified 25 food groups associated with dementia in a food-wide association analysis. Second, we ranked their importance using machine learning and prioritized eight groups (for example, green leafy vegetables, berries and citrus fruits). Finally, we established and externally validated a MODERN score (0–7), which showed stronger associations with lower risk of dementia-related outcomes (hazard ratio comparing highest versus lowest tertiles: 0.64, 95% CI: 0.43–0.93) than the a priori-defined MIND diet (0.75, 0.61–0.92). Across 63 health-related outcomes, the MODERN diet showed particularly significant associations with mental/behavioural disorders. Multimodal neuroimaging, metabolomics, inflammation and proteomics analyses revealed potential pathways and further support the potential of MODERN diet for dementia prevention.


  • Open Access

    Temporal trends of blood-based markers in various psychiatric disorders and their cross-sectional brain structure associations

    by Yu-Jia Wang, Zairen Zhou, Yu-Zhu Li, Ju-Jiao Kang, Jin-Tai Yu, Jian-Feng Feng, Wei Cheng, Guoqing Pan, Jia You, Linbo Wang

    Background: Understanding the temporal trends of blood-based biomarkers and their associations with brain structure is crucial for early detection and intervention in psychiatric disorders. This study aimed to explore these trends in the decade before and after diagnosis, along with the cross-sectional relationships with brain structures. Methods: Utilizing UK Biobank data, we conducted a nested case-control analysis of individuals aged 40-69 years diagnosed with anxiety (n = 27,216), bipolar disorder (n = 1325), depression (n = 36,570), or schizophrenia (n = 1478) within 10 years of baseline. We used multivariable linear regression to analyze temporal trends and brain structure associations for 31 blood cell counts, 28 biochemistry markers, and 168 serum metabolites. Results: Here we show that compared to controls, significant temporal divergence is observed in 39, 6, 55, and 12 blood-based markers for anxiety, bipolar disorder, depression, and schizophrenia, respectively. Common biomarkers like cystatin C, red blood cells, hemoglobin, hematocrit, and total bilirubin are identified. Biomarkers cluster into groups with either linear or non-linear trends. Among the linearly changing biomarkers, some have a widening difference from controls while others have a narrowing one. For example, in the case of depression, HDL-TG demonstrates an increasing disparity over time, while cholesterol exhibits a decreasing trend. Non-linear clusters often show reversals around diagnosis, indicating potential treatment effects. Differential associations are found between biomarkers and brain regions, including the orbitofrontal cortex, hippocampus, and accumbens. Conclusions: This study reveals the temporal trends of blood-based biomarkers in psychiatric disorders and their correlations with brain structure, aiding early detection and potentially enhancing clinical outcomes.


  • Open Access

    Atlas of proteomic signatures of brain structure and its links to brain disorders

    by Peng Ren, Xiao-He Hou, Zeyu Li, Jia You, Yuzhu Li, Wei Zhang, Weikang Gong, Bei Zhang, Bangsheng Wu, Linbo Wang, et al.

    Individual variation in brain structure influences deterioration due to disease and comprehensive profiling of the associated proteomic signature advances mechanistic understanding. Here, using data from 4997 UK Biobank participants, we analyzed the associations between 2920 plasma proteins and 272 neuroimaging-derived brain structure measures. We identified 5358 associations between 1143 proteins and 256 brain structure measures, with NCAN and LEP proteins showing the most associations. Functional enrichment implicated these proteins in neurogenesis, immune/apoptotic processes and neurons. Furthermore, bidirectional Mendelian randomization revealed 33 associations between 32 proteins and 23 brain structure measures, and 21 associations between nine brain structure associated proteins and ten brain disorders. Moreover, the significant associations between the identified proteins and mental health were mediated by brain volume and surface area. In summary, this study generates a comprehensive atlas mapping the patterns of association between proteome and brain structure, highlighting their potential value for studying brain disorders.


  • Top Cited Papers

    Large-scale proteomic analyses of incident Parkinson’s disease reveal new pathophysiological insights and potential biomarkers

    by Yi-Han Gan, Ling-Zhi Ma, Yi Zhang, Jia You, Yu Guo, Yu He, Lin-Bo Wang, Xiao-Yu He, Yu-Zhu Li, Qiang Dong, et al.

    The early pathophysiology of Parkinson’s disease (PD) is poorly understood. We analyzed 2,920 Olink-measured plasma proteins in 51,804 UK Biobank participants, identifying 859 incident PD cases after 14.45 years. We found 38 PD-related proteins, with six of the top ten validated in the Parkinson’s Progression Markers Initiative (PPMI) cohort. ITGAV, HNMT and ITGAM showed consistent significant association (hazard ratio: 0.11–0.57, P = 6.90 × 10−24 to 2.10 × 10−11). Lipid metabolism dysfunction was evident 15 years before PD onset, and levels of BAG3, HPGDS, ITGAV and PEPD continuously decreased before diagnosis. These proteins were linked to prodromal symptoms and brain measures. Mendelian randomization suggested ITGAM and EGFR as potential causes of PD. A predictive model using machine learning combined the top 16 proteins and demographics, achieving high accuracy for 5-year (area under the curve (AUC) = 0.887) and over-5-year PD prediction (AUC = 0.816), outperforming demographic-only models. It was externally validated in PPMI (AUC = 0.802). Our findings reveal early peripheral pathophysiological changes in PD crucial for developing early biomarkers and precision therapies.