SISG INTEGRATIVE GENOMICS MODULE IG3

This module emphasizes how the theory and application of transcriptomics can be extended to include other types of omic analysis, and then integrated using statistical and machine learning tools. It starts with the statistical basis of hypothesis testing, covering the central role of normalization strategies and the specifics of differential expression analysis. The module then discusses options for downstream processing by clustering and module comparison; extensions to proteomics, and metabolomics; eQTL analysis including fine mapping of regulatory variation; and integrative methodologies addressing the relationship between genomic, meta-genomic, and phenotypic variation. The module will also emphasize the landscape of computational problems in single-cell RNA sequencing, including clustering, trajectory inference, imputation, and data integration. This module deals primarily with upstream data processing methods that lead to the delineation of networks and pathways that are discussed further in Module 18. Two sessions are devoted to worked exercises.

Learning Objectives: After attending this module, participants will be able to: 

1. Appreciate how good experimental design is central to being able to extract more inference from gene expression profiling experiments.

2. Understand the importance of normalization strategies to avoid biases and maximize statistical power to detect biological effects.

3. Implement open source DESeq or edgeR that is commonly used for transcriptomic hypothesis testing.

4. Understand the principles of eQTL analysis and the computational methods for dissecting regulatory mechanisms by integration with chromatin profiling data.  

5. Have a good understanding of the landscape of computational problems in single-cell RNA sequencing analysis, including the theory of cluster analysis and trajectory inference.

Saumya Jain is Assistant Professor in the School of Biological Sciences at Georgia Tech. He is fascinated by the assembly of brains consisting of hundreds of millions of neurons whose spatial and temporal differentiation and connectity is orchestrated by gene expression programs.  To this end, his lab delpoys cutting edge single nucleus transcriptomic and spatial transcriptomic technologies in flies and mice. Learn more about Saumya’s work here.

Peng Qiu is an Associate Professor of Biomedical Engineering at Georgia Tech.  His research is focused on Machine Learning and Bioinformatics, by developing novel computational methods for advancing biological discoveries. Current research projects include machine learning analysis on single-cell data, multi-omics integration in cancer, experimental design and model reduction in systems biology. Learn more about Peng’s work here.