Integrating big data and computational algorithms to decipher human gut microbiome with Chinese Medicine
In the last decade, human gut microbiome becomes hotspot in precision medicine. It has been proved to be associated with many human complex diseases and affect the effectiveness of Chinese medicine. The dynamic change of microbial composition and abundance prompt its wide applications in exploring disease biomarkers and drug targets. However, the current studies only focus on the gene level, the involved microbes have not been deeply investigated yet. In this project, we plan to build a comprehensive global human gut metagenome catalog and develop novel computational algorithms and machine learning models to facilitate downstream data analysis. We believe this project will become a milestone in metagenome research and open a door for its application in precision Chinese medicine.
Aims of the project
- Integrate big data to construct reference human gut microbiota catalog.
We propose to collect and sequence ~10,000 individuals from diverse populations, lifestyles, disease status and ages to integrate their sequencing data into a uniform human gut microbiota catalog using graphical models.
- Develop novel computational approached to handle emerging sequencing technologies.
Over the past decades, cost-effective whole-genome sequencing has been revolutionized by short-fragment approaches. Currently, two alternative approaches are offered by linked-reads and long-reads. But there lacks efficient software to handle these recently emergent technologies and make full use of DNA long-range information. We aim to develop a series of computational algorithms and tools to analyze multi-platform metagenome sequencing data.
- Design machine learning models to infer enterotype-phenotype relationships.
The genotype-phenotype relationships has been well studied based on the genomic variants from human genome. Such information could facilitate our understanding of disease pathology and can even be used to single out the disease high risk individuals from general population. Recent studies realized the individuals could be classified based on the composition of gut microbiome, called enterotype. Different from genotype, enterotype is dynamic and determined by both internal and external factors. We sought to develop novel machine learning and deep learning models to infer the connections between enterotype-phenotype and further explore microbial biomarkers for both disease and Chinese Medicine.
The candidates require extensive background in algorithms, machine learning or bioinformatics, with experience in relevant programming languages (e.g. C++, Python).