Le Zhang

Le Zhang

Auto-generated profile, based on publicly available information

Le Zhang

Center for Multimodal Imaging and Genetics, J. Craig Venter Institute, La Jolla, CA, USA

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 March 2025 in Journal of Medical Internet Research
Verified
Journal Article
Open access
RHRosemary He
Rosemary He
VSVaruni Sarwal
Varuni Sarwal
XQXinru Qiu
Xinru Qiu
YZYongwen Zhuang
Yongwen Zhuang
LZLe Zhang
Le Zhang
JCJeffrey Chiang
Jeffrey Chiang
Background: Trajectory modeling is a long-standing challenge in the application of computational methods to health care. In the age of big data, traditional statistical and machine learning methods do not achieve satisfactory results as they often fail to capture the complex underlying distributions of multimodal health data and long-term dependencies throughout medical histories. Recent advances in generative artificial intelligence (AI) have provided powerful tools to represent complex distributions and patterns with minimal underlying assumptions, with major impact in fields such as finance and environmental sciences, prompting researchers to apply these methods for disease modeling in health care. Objective: While AI methods have proven powerful, their application in clinical practice remains limited due to their highly complex nature. The proliferation of AI algorithms also poses a significant challenge for nondevelopers to track and incorporate these advances into clinical research and application. In this paper, we introduce basic concepts in generative AI and discuss current algorithms and how they can be applied to health care for practitioners with little background in computer science. Methods: We surveyed peer-reviewed papers on generative AI models with specific applications to time-series health data. Our search included single- and multimodal generative AI models that operated over structured and unstructured data, physiological waveforms, medical imaging, and multi-omics data. We introduce current generative AI methods, review their applications, and discuss their limitations and future directions in each data modality. Results: We followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines and reviewed 155 articles on generative AI applications to time-series health care data across modalities. Furthermore, we offer a systematic framework for clinicians to easily identify suitable AI methods for their data and task at hand. Conclusions: We reviewed and critiqued existing applications of generative AI to time-series health data with the aim of bridging the gap between computational methods and clinical application. We also identified the shortcomings of existing approaches and highlighted recent advances in generative AI that represent promising directions for health care modeling.

11 May 2016 in Birth Defects Research Part A: Clinical and Molecular Teratology
Verified
Journal Article
Open access
PDF file
NYNa Yang
Na Yang
LWLinlin Wang
Linlin Wang
RFRichard H. Finnell
Richard H. Finnell
ZLZhiwen Li
Zhiwen Li
LJLei Jin
Lei Jin
ARAiguo Ren
Aiguo Ren
Background: After years of periconceptional folic acid supplementation, the prevalence of neural tube defects (NTDs) remains stable following the remarkable reduction observed immediately after the fortification practice. There is accumulating evidence that folate receptor (FR) autoimmunity may play a role in the etiology of folate‐sensitive NTDs.Methods: From 2011 to 2013, 118 NTD cases and 242 healthy controls were recruited from a population‐based birth defects surveillance system in Northern China. Enzyme‐linked immunosorbent assay was used to measure FR autoantibodies in maternal and cord blood. Logistic regression models were used to estimate the odds ratios (OR) and 95% confidence intervals (95% CI).Results: Plasma FR autoantibodies levels were significantly elevated in mothers of infants with NTDs compared with mothers of healthy controls. Using the lowest tertile as the referent group, 2.20‐fold (95% CI, 0.71–6.80) and 5.53‐fold increased odds (95% CI, 1.90–16.08) of NTDs were observed for the second and third tertile of immunoglobulin G (IgG), respectively, and the odds of NTDs for each successive tertile of IgM was 0.98 (95% CI, 0.35–2.75) and 3.49 (95% CI, 1.45–8.39), respectively. A dose–response relationship was found between FR autoantibodies levels and risk of NTDs (P< 0.001 for IgG,P= 0.002 for IgM). The same pattern was observed in both subtypes of spina bifida and anencephaly. No significant difference in levels of cord blood FR autoantibodies was observed.Conclusion: Higher levels of FR autoimmunity in maternal plasma are associated with elevated risk of NTDs in a dose–response manner. Birth Defects Research (Part A) 106:685–695, 2016.

1 January 2022 in Biochemical and Biophysical Research Communications
Verified
Journal Article
Open access
NGNingning Guo
Ningning Guo
LZLe Zhang
Le Zhang
WFWei Fan
Wei Fan
LBLi-Ping Bai
Li-Ping Bai
XZXianwen Zhang
Xianwen Zhang
JBJie Bai
Jie Bai
Morphine is the pain releasing and abusing drug. Morphine leads to addiction by activating dopaminergic rewarding system consisted of the ventral tegmental area (VTA) and nucleus accumbens (NAc). Cholecystokinin (CCK) is a gut-brain neuropeptide and involved in morphine dependence. Brain-derived neurotrophic factor (BDNF) is a neurotrophin and plays roles in regulating addiction. Geranylgeranylacetone (GGA) is a medicine of protecting gastric mucosal injury and protecting neurons. Our previous study showed that GGA blocked morphine-induced withdrawal and relapse through inducing thioredoxin 1(Trx1). In this study, we investigated that whether cholecystokinin-B receptor (CCKB receptor) and BDNF were related to GGA inhibition on morphine addiction. At first, we made conditioned place preference (CPP) model and confirmed again that GGA blocked the expression of morphine-CPP in present study. Then, our results showed that morphine increased the expressions of dopamine D1 receptor, tyrosine hydroxylase (TH), CCKB receptor and BDNF in the VTA and NAc in mice, which was inhibited by GGA. These results suggest that CCK and BDNF in dopaminergic systems are associated with the role of GGA blocking morphine-CPP.

9 September 2025 in bioRxiv
Verified
Preprint
Open access
LZLe Zhang
Le Zhang
OCOlivier Celhay
Olivier Celhay
BDBiplabendu Das
Biplabendu Das
SBSamantha Berman
Samantha Berman
LZLaura Ziemer
Laura Ziemer
JLJanosch Linkersdörfer
Janosch Linkersdörfer
The ABCD Study® is the largest longitudinal research study of brain development and child health across the United States. The study generates a rich multimodal dataset, including file-based raw and derivative data and a curated tabulated dataset comprising data from a broad spectrum of research domains. Historically, summary scores in the tabulated dataset were computed using the electronic data capture system REDCap or external scripting solutions. However, this approach was rigid, unreliable, and lacked transparency around how scores were computed. Here, we introduce ABCDscores, an R software package that is used by the study to compute most non-proprietary summary scores included in the tabulated ABCD data resource and allows for customization to meet specific research needs. ABCDscores provides a flexible, open-source solution that uses R for transparent and reproducible scoring. The package is built to apply standard scoring methods across different types of assessments, ensuring consistent and accurate results even with large datasets. By simplifying complex scoring procedures into a user-friendly process, ABCDscores streamlines research workflows, reduces errors, and saves researchers valuable time. The package aims to improve analyses of data from the ABCD Study® and can serve as a model for the development of similar tools and a standardized framework for other research studies, fostering new discoveries and innovation in the field.

30 April 2024 in JMIR Preprints
Verified
Preprint
Open access
RHRosemary He
Rosemary He
VSVaruni Sarwal
Varuni Sarwal
XQXinru Qiu
Xinru Qiu
YZYongwen Zhuang
Yongwen Zhuang
LZLe Zhang
Le Zhang
JCJeffrey Chiang
Jeffrey Chiang
BACKGROUND Trajectory modeling is a long-standing challenge in the application of computational methods to health care. In the age of big data, traditional statistical and machine learning methods do not achieve satisfactory results as they often fail to capture the complex underlying distributions of multimodal health data and long-term dependencies throughout medical histories. Recent advances in generative artificial intelligence (AI) have provided powerful tools to represent complex distributions and patterns with minimal underlying assumptions, with major impact in fields such as finance and environmental sciences, prompting researchers to apply these methods for disease modeling in health care. OBJECTIVE While AI methods have proven powerful, their application in clinical practice remains limited due to their highly complex nature. The proliferation of AI algorithms also poses a significant challenge for nondevelopers to track and incorporate these advances into clinical research and application. In this paper, we introduce basic concepts in generative AI and discuss current algorithms and how they can be applied to health care for practitioners with little background in computer science. METHODS We surveyed peer-reviewed papers on generative AI models with specific applications to time-series health data. Our search included single- and multimodal generative AI models that operated over structured and unstructured data, physiological waveforms, medical imaging, and multi-omics data. We introduce current generative AI methods, review their applications, and discuss their limitations and future directions in each data modality. RESULTS We followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) guidelines and reviewed 155 articles on generative AI applications to time-series health care data across modalities. Furthermore, we offer a systematic framework for clinicians to easily identify suitable AI methods for their data and task at hand. CONCLUSIONS We reviewed and critiqued existing applications of generative AI to time-series health data with the aim of bridging the gap between computational methods and clinical application. We also identified the shortcomings of existing approaches and highlighted recent advances in generative AI that represent promising directions for health care modeling.

4 November 2026 in bioRxiv
Verified
Preprint
Open access
PDF file
LZLe Zhang
Le Zhang
BDBiplabendu Das
Biplabendu Das
OCOlivier Celhay
Olivier Celhay
CSCalen J. Smith
Calen J. Smith
LZLaura R. Ziemer
Laura R. Ziemer
JLJanosch Linkersdörfer
Janosch Linkersdörfer
The NIH Brain Development Cohorts (NBDC) Data Hub provides access to large-scale longitudinal datasets, including the Adolescent Brain Cognitive Development (ABCD) Study® and the HEALthy Brain and Child Development (HBCD) Study. The studies release their tabulated data as hundreds of tables with tens of thousands of variables. Turning these files into a dataset for a particular analysis poses substantial informatics challenges, especially when data preparation and analysis workflows must be reproducible and reusable for new releases. We therefore developed NBDC Data Tools , a suite of open-source packages for preparing NBDC data in local or other approved computing environments. NBDCtools is the core R package for dataset construction and processing. The Python package nbdctools either calls the R workflow or assembles datasets natively in Python, and the R package NBDCtoolsData provides the release metadata support for both NBDCtools and nbdctools . Users define a dataset in code by naming the study, the release, and the variables or tables they need. The software then uses the standardized metadata released by the studies to locate the relevant files, join tables on the participant and session identifiers, convert data types, attach labels, and optionally recode non-response codes as missing values. Because the dataset definition is a short script that runs on the original release files, it can be rerun, shared with collaborators, and reported in publications. NBDC Data Tools complements the web-based data platforms of the NBDC Data Hub by providing a reproducible, scriptable path from approved data access to analysis-ready data for statistical analysis.