This research topic focuses on analyzing the genomic changes in pancreatic cancer cells and circulating free DNA (cfDNA) in patients' blood. High-throughput sequencing technology is used to identify key mutations including single nucleotide polymorphisms (SNPs), copy number variations (CNVs) and structural variations, hoping to discover molecular markers with clinical potential to assist in early diagnosis and prognosis assessment.

This research direction integrates RNA-seq and miRNA-seq and other multi-omics data, combined with machine learning algorithms (such as gradient boosting trees), to select features and predict risks for key genes and signal pathways that are potential cancers in colorectal polyps. Through large-scale data mining and model training, a risk prediction tool for personalized medical applications is established.

This research topic focuses on centriolar proteins that are abnormally expressed in specific cancers. Dimensionality reduction techniques such as t-Stochastic Neighbor Embedding (t-SNE) are used to screen proteins that are highly correlated with the pathological development of cancer, and further explore the relationship between their functions and biological processes such as cancer cell mitosis and chromosomal instability.

We isolated human mesenchymal stem cells from placenta and established an in vitro model of their differentiation into adipocytes and neurons. We also analyzed the regulatory role of exosomes secreted by them in the physiological mechanism of stem cells and their potential in regenerative medicine and clinical applications.