About

Hua Tang is a Professor of Genetics in the School of Medicine at Stanford University and also holds the courtesy title of Professor of Statistics. He earned his graduation in 2002 with a dissertation titled 'Statistical Methods for Two Problems in DNA Sequence Comparisons' under the advisement of Siegmund. His research interests include statistical genetics, computational biology, and population genomics. He is associated with the Tang Lab in the Department of Genetics and has contributed to the development of statistical methods for DNA sequence analysis, focusing on genetic data and biological computations.

Research topics

• Genetics
• Biology
• Evolutionary biology
• Computational biology
• Medicine
• Endocrinology
• Microbiology
• Immunology
• Demography
• Internal medicine
• Cell biology
• Pathology
• Bioinformatics

Selected publications

• Prediction of bowel resection due to bowel necrosis in recent portal vein thrombosis based on CT presentations and clinical data

  BMC Gastroenterology · 2025-11-27

  articleOpen access

  BACKGROUND: This study aims to identify predictors of bowel resection due to bowel necrosis in patients with recent portal vein thrombosis (PVT) based on computed tomography (CT) presentations and clinical data. METHODS: This retrospective study was conducted from January 2013 to June 2023 and included patients with noncirrhotic, nontumoral recent PVT. Patients were followed up from initial hospitalization until September 2023 or death. Presentations on CT images, laboratory tests, adverse events, and clinical outcomes were recorded. Independent predictors of bowel resection due to bowel necrosis were identified using univariate and multivariate logistic regression models. RESULTS: Seventy-six patients (mean age, 41.6 years ± 13.0; 57 males) were evaluated. All patients received low molecular weight heparin, fasting, and intravenous antibiotics during hospitalization, and long-term oral anticoagulation after discharge. CT imaging revealed completely occlusive main portal vein in 73.7% (56/76) of patients and partially occlusive in 26.3% (20/76). Thrombus extension into the superior mesenteric vein was present in 80.3% (61/76) of patients and into the splenic vein in 69.7% (53/76). Additionally, 44.7% (34/76) of patients were classified as high density (CT value of main portal vein ≥ 50 Hounsfield Units on unenhanced CT scans), while 55.3% (42/76) were classified as low density (CT value < 50 Hounsfield Units). Bowel resection was performed in 15.8% (12/76) of patients due to bowel necrosis. Independent predictors of bowel resection were low CT value of the portal vein thrombus and a high neutrophil count. CONCLUSIONS: In patients with noncirrhotic, nontumoral recent PVT, a low CT value of the main portal vein thrombus and a high neutrophil count are important predictive factors for bowel resection due to bowel necrosis.

  PublisherDOI

• Synthesis of an autochthonous microbiota responsible for the prepared process improving, flavor profile and brewing attributes of high-temperature Daqu

  Food Bioscience · 2025-03-29 · 1 citations

  article
  PublisherDOI

• Robust inference with GhostKnockoffs in genome-wide association studies

  Research Square · 2025-05-05

  preprintOpen accessSenior author
  PublisherOA PDFDOI

• rSIG combined with NLR in the prognostic assessment of patients with multiple injuries

  Open Medicine · 2025-01-01

  articleOpen accessSenior author

  Abstract Objective To investigate the significance of the reverse shock index multiplied by the Glasgow Coma Scale score (rSIG) and the neutrophil–lymphocyte ratio (NLR) in the prediction of prognosis in patients with multiple injuries. Methods The clinical data of 142 patients with multiple injuries admitted to the Emergency Department of Shenzhen Longhua District Central Hospital between January 2019 and December 2022 were retrospectively analyzed. Subjects were divided into the survival group ( n = 102) and the deceased group ( n = 40) based on their survival status at 28 days after injury. We subsequently analyzed the intergroup differences in blood test results, rSIG, and NLR, as well as the relationship between rSIG and NLR. The predictive value of rSIG, NLR, and both combined in determining the prognosis of patients with multiple injuries was explored by plotting the receiver operator characteristic (ROC) curve. Based on the optimal cut-point value of the ROC curves, subjects were divided into groups with rSIG ≤ 7.75 (22 patients) and rSIG &gt; 7.75 (120 patients), as well as groups with NLR ≤ 10.36 (104 patients) and NLR &gt; 10.36 (38 patients), and the 28-day mortality rate was compared between the groups. Results A total of 142 patients were enrolled. The rSIG of the survival group ( n = 102) was significantly greater (15.7 ± 4.8) than that of the deceased group ( n = 40, 6.2 ± 2.9), ( t = 14.307, p &lt; 0.001). The NLR of the survival group was higher than that of the deceased group, but the difference was not statistically significant ( p &gt; 0.05). The area under the curve (AUC) of the ROC of NLR was significantly greater than that of rSIG (0.922 vs 0.54) ( Z = −7.881, p &lt; 0.001). The AUC for NLR was also statistically greater than that of the combination of rSIG and NLR (0.963 vs 0.54) ( Z = −8.378, p &lt; 0.001). The AUC of the combination of rSIG and NLR was significantly greater than that of rSIG (0.844 vs 0.540) ( Z = 2.239, p = 0.025). The 28-day mortality rate of patients in the rSIG ≤ 7.75 group was also significantly greater than that of patients in the rSIG &gt; 7.75 group (10.0%) ( p &lt; 0.05). Finally, the 28-day mortality rate in the group with an NLR ≤ 10.36 was lower than that in the group with an NLR &gt; 10.36 ( p &lt; 0.05). Pearson correlation analysis showed that the correlation coefficient between rSIG and NLR was r = 0.13, which did not reach statistical significance ( p = 0.12). Conclusion NLR, rSIG, and the combination of the two are all valuable in predicting the prognosis of patients with multiple injuries (all AUC &gt; 0.5). However, the predictive capacity of NLR was better than either rSIG alone or both combined. These findings may serve as references in guiding the treatment of patients with multiple injuries in clinical practice.

  PublisherOA PDFDOI

• Clinical efficacy of combined detection of serum procalcitonin and C-reactive protein in early differential diagnosis of bacterial and viral pneumonia and analysis of related inflammatory response mechanisms

  Biomarkers in Medicine · 2025-07-31 · 3 citations

  articleSenior authorCorresponding

  BACKGROUND: Early distinction of bacterial etiology from viral causes represents a cornerstone for rational antimicrobial therapy. Through evaluation of dual-biomarker strategies, this investigation examined the diagnostic utility of simultaneous PCT and CRP measurement while elucidating underlying inflammatory pathways. METHODS: = 122) hospitalized during January 2023 through December 2024. Laboratory parameters including PCT, CRP, leukocyte counts, neutrophil proportions, IL-6, and TNF-α underwent comprehensive measurement. Diagnostic accuracy was evaluated through ROC analysis, with severity correlations systematically assessed. RESULTS: < 0.001). CONCLUSION: The synergistic assessment of PCT with CRP surpasses individual marker performance for bacterial-viral pneumonia discrimination, capturing differential inflammatory cascades and severity stratification. This dual-biomarker strategy optimizes therapeutic decisions and antimicrobial governance.

  PublisherDOI

• ProMix: Enhancing Protein Quantification through Experimental Design and Statistical Normalization

  Journal of Proteome Research · 2025-07-07

  articleSenior authorCorresponding

  Isobaric labeling of biospecimens followed by mass spectrometry (MS) has become the method of choice for large-scale, untargeted, quantitative proteomic profiling. However, subtle variation in experimental conditions can amplify sample variability and introduce systematic biases. Motivated by the challenges and opportunities arose in a recent proteogenomic study, we developed ProMix, a flexible analytical framework designed to improve protein normalization by leveraging two key experimental design features: (1) the inclusion of an additional reference sample to serve as an internal standard, and (2) the incorporation of replicates of each specimen. ProMix can utilize either or both features. Through applications to both simulated and real data sets, we demonstrate the improved performance of ProMix and highlight the advantages of the enhanced experimental design strategies.

  PublisherDOI

• Physicochemical Properties, Flavor and Microbial Community of Clean Low-Temperature Daqu Originated from Synthetic Autochthonous Microbiota

  EAS Journal of Nutrition and Food Sciences · 2025-01-06

  articleOpen access1st authorCorresponding

  Based on the 23 strains of low-temperature Daqu, microbial inoculants were prepared and inoculated into crushed barley and pea to prepare clean low-temperature Daqu (XQ). The physicochemical properties, flavor profile, and microbial community of XQ were investigated, with traditional low-temperature Daqu (CQ) and production requirements as controls. The results indicated that there was no significant difference in moisture and acidity between the two types of Daqu: CQ exhibited a moisture of 10.6% and an acidity of 1.1 mmol/10 g, while XQ demonstrated a moisture of 10.9% and an acidity of 1.2 mmol/10 g. The fermenting activity (1.36 g/0.5 g·72 h) and liquefying activity (1.13 g/g·h) of XQ surpassed those of CQ; however, its saccharifying activity (780 mg/g·h) and esterifying activity (808 mg/50 g·7 d) were lower. In general, the physicochemical properties of XQ align with the production requirements. HS-SPME-GC-MS analysis indicated that the flavor profiles of the two types of Daqu were largely similar, with 84.85% of the flavor components of CQ being reproducible in XQ. Microbiota community analysis revealed there were some differences in relative abundance of microbes and COG (Clusters of Orthologous Groups) function between the two Daqu. The dominant microorganisms in XQ identified were Bacillus, Pichia, Transversalis, and Monascus, meanwhile the dominant microorganisms in CQ identified were Pediococcus, Rhizomucor, Wickerhamomyces. This study establishes a robust experimental basis for producing clean Daqu and provides valuable insights for the development of safer microbial fermented foods.

  PublisherDOI

• Expectations for papers performing Mendelian randomization analyses

  PLoS Genetics · 2025-07-17 · 4 citations

  editorialOpen access
  PublisherDOI

• Beyond guilty by association at scale: searching for causal variants on the basis of genome-wide summary statistics

  bioRxiv (Cold Spring Harbor Laboratory) · 2024-03-03 · 5 citations

  preprintOpen access

  Understanding the causal genetic architecture of complex phenotypes will fuel future research into disease mechanisms and potential therapies. Here, we illustrate the power of a novel framework: it detects, starting from summary statistics, and across the entire genome, sets of variants that carry non-redundant information on the phenotypes and are therefore more likely to be causal in a biological sense. The approach, implemented in open-source software, is also computationally efficient, requiring less than 15 minutes on a single CPU to perform genome-wide analysis. Through extensive genome-wide simulation studies, we show that the method can substantially outperform existing methods in false discovery rate control, statistical power and various fine-mapping criteria. In applications to a meta-analysis of ten large-scale genetic studies of Alzheimer's disease (AD), we identified 82 loci associated with AD, including 37 additional loci missed by conventional GWAS pipeline. Massively parallel reporter assays and CRISPR-Cas9 experiments have confirmed the functionality of the putative causal variants our method points to. Finally, we retrospectively analyzed summary statistics from 67 large-scale GWAS for a variety of phenotypes. Results reveal the method's capacity to robustly discover additional loci for polygenic traits and pinpoint potential causal variants underpinning each locus beyond conventional GWAS pipeline, contributing to a deeper understanding of complex genetic architectures in post-GWAS analyses.

  PublisherOA PDFDOI

• Author Reply to Peer Reviews of Recruitment of the m6A/Am demethylase FTO to target RNAs by the telomeric zinc finger protein ZBTB48

  2024-06-01

  peer-review
  PublisherDOI

Recent grants

• NIH Grant U01HG007611

  NIH · $235k · 2017

• NIH Grant R01HG007078

  NIH · $886k · 2017

• NIH Grant R01GM073059

  NIH · $2.4M · 2018

• Machine Learning for Decoding the Genetic Architecture of Complex Traits and Diseases Across Populations

  NIH · $2.2M · 2018–2031

• NIH Grant U01HG00761101

  NIH · $1.5M · 2017

Frequent coauthors

• Jun Liu

  Suzhou University of Science and Technology

  864 shared

• Wei Zhao

  Michigan United

  748 shared

• Zhe Wang

  Zhejiang University

  576 shared

• Wei Zhou

  XinHua Hospital

  460 shared

• Wei Zhou

  Yanbian University

  444 shared

• Jun Liu

  University of California, San Francisco

  441 shared

• Xiaofeng Zhu

  Case Western Reserve University

  359 shared

• Wei Zheng

  340 shared