UT Southwestern Medical Center


Guanghua Xiao is a Professor (with Tenure) in the Department of Population & Data Sciences and the Department of Bioinformatics. He is appointed as the Mary Dees McDermott Hicks Chair in Medical Science. He is also a founding member of the Quantitative Biomedical Research Center at UT Southwestern Medical Center. He received his bachelor's training in Engineering from Tsinghua University and a Ph.D. degree in Biostatistics from the University of Minnesota. Dr. Xiao is the PI of a CPRIT (Cancer Prevention Institute of Texas) grant focusing on analyzing digital pathology data to improve lung cancer patient care. He was also the PI of an NIH R01 grant focusing on developing novel computational models to analyze genomics data to identify new potential therapeutic targets of lung cancer, and the PI on an NIH R21/R33 grant on developing integrative analysis methods to study genome-wide mRNA expression and epigenomic data. Dr. Xiao has developed novel computational models for integrative analysis, and he has extensive experience in the analysis of high throughput data such as next generation sequencing (NGS) data and tissue imaging data. His team has developed many user-friendly software and tools to analyze high-dimensional and complex genomic and imaging data. In recent years, his lab is focusing on developing AI methods for tissue imaging analysis and bioinformatics tools for spatial molecular profiling technologies.

Read More

Research Summary

I am interested in developing computational models and algorithms for big data to predict patients' outcomes, which can help clinicians to tailor treatment plans for individual patients. My research lab has developed statistical methodologies for Bayesian analysis, new bioinformatics tools, and gene expression signatures to predict patient prognosis and response to chemotherapy. We are actively developing new bioinformatics tools and computational algorithms for big data, such as imaging and genomic data. Currently, my research is mainly focused on tissue image analysis. Our team was among the first teams to develop and validate computational models using histopathology images collected from routine clinical procedures to refine lung cancer prognosis (Luo et al, 2017, Luo et al, 2019). We have developed deep learning-based models to detect tumor regions, micro-blood-vessels and predict patient outcomes (Wang et al, 2018; Yi et al, 2018; Huang et al, 2017). Recently, we developed algorithms to detect and classify different types of cells from tissue images (Wang et al, 2019; Wang et al, 2020) and developed a set of spatial models (Li et al, 2019, Li et al, 2020) to investigate cell spatial organizations and their implication in disease. Aside from my research in developing statistical methodology, I am also devoted to applying state-of-the-art statistical methods in biomedical research.

Research Interests

  • Tissue imaging data analysis
  • Machine learning and deep learning
  • Developing computational algorithms and bioinformatics tools for complex biomedical data
  • Developing spatial models for biological data
  • Biomedical imaging analysis

Selected Publications

Based on an analysis of two large databases on breast cancer, reduced activity of an autophagy gene, beclin 1, was related to both a higher incidence of triple-negative breast cancer and a poorer prognosis for breast cancer patients.

Genomic regression analysis of coordinated expression.

Nature Communications 8, Article number: 2187 (2017) doi:10.1038/s41467-017-02181-0
Cai L, Li Q, Du Y, Yun J, Xie Y, DeBerardinis RJ, Xiao G.

publisher's website

Using functional signature ontology (FUSION) to identify mechanisms of action for natural products.

Science Signal, Vol. 6, Issue 297, pp. ra90
Potts MB, Kim HS, Fisher KW, Hu Y, Carrasco YP, Bulut GB, Ou YH, Herrera-Herrera ML, Cubillos F, Mendiratta S, Xiao G, Hofree M, Ideker T, Xie Y, Huang LJ, Lewis RE, MacMillan JB, White MA.

publisher's website

Akt-mediated regulation of autophagy and tumorigenesis through Beclin 1 phosphorylation.

Science, Vol. 338, Issue 6109, pp. 956-959 DOI: 10.1126/science.1225967
Wang RC, Wei Y, An Z, Zou Z, Xiao G, Bhagat G, White M, Reichelt J, Levine B.

publisher's website

Try Out Our Software

We have developed online analysis tools which allow users to explore and analyze lung cancer- and germ cell tumor-related gene expression data. PIPECLIP Galaxy is also provided here for biologists to identify the most likely cross-linking sites.