Back to

main Publications Talks Graduate Students Teaching Links

My research focuses on the development of statistical and machine learning methodologies for high-dimensional biological data, with key contributions in the following areas:

---

1. Statistical Modeling, Inference, and Multisource Omics Data

· Develop methods for integrating heterogeneous omics data (genomics, microbiome, clinical)

· Address challenges of high dimensionality, noise, and measurement error

· Contributions include:

o Deconvolution methods for measurement error correction

o Non-negative matrix factorization (NMF) for latent structure discovery

o Cross-study integration methods for combining multiple datasets

---

2. Computational Molecular Evolution

· Develop statistical models for evolutionary processes and phylogenetics

· Key contributions:

o Codon substitution models and likelihood-based clustering (LiBaC)

o Methods for testing model adequacy and selection pressure inference

o Gene coevolution and phylogenetic modelling

---

3. Dimension Reduction, Variable Selection, and FDR Control

· Advance methods for high-dimensional data analysis

· Contributions include:

o Dimension reduction (e.g., Poisson PCA, interpretable PCA)

o Sparse variable selection methods (e.g., SuRF)

o False discovery rate (FDR) control, including hierarchical and structured approaches

· Emphasis on scalability, theoretical guarantees, and interpretability

---

4. Machine Learning Methods and Applications in Medicine

· Develop machine learning and AI methods for healthcare applications

· Focus areas:

o Neural network-based feature and structure selection

o Clinical decision support systems and predictive modelling

o Applications in medical imaging, diagnostics, and emergency medicine

· Emphasis on interpretable and reliable AI models

---

5. Temporal Dynamics of the Microbiome

· Model time-evolving microbial systems using statistical and stochastic approaches

· Contributions include:

o Stochastic differential equation models (e.g., Ornstein–Uhlenbeck processes)

o Optimal sampling design for time-series data

o Integration with dimension reduction and variable selection methods

· Applications to understanding microbial community dynamics and health outcomes

---

Overall Research Vision

· Integrate statistical theory and machine learning into unified frameworks

· Develop robust, interpretable, and scalable methods for complex biological data

· Advance applications in biomedical research, systems biology, and precision medicine