BrainGene
Gene Expression Analysis for Vertebrate Brain Development
Overview
The BrainGene project developed a DNA chip-based gene expression analysis system targeting approximately 10,000 rat genes (out of ~30,000 total), with the goal of identifying genes specifically expressed during vertebrate brain development. The project combined experimental molecular biology with advanced computational machine learning to profile gene expression across developmental stages and brain regions.
A central outcome was the construction of a gene expression database with a web interface, making the curated results accessible to the broader neuroscience and bioinformatics community.
Research Objectives
- Design and fabricate a DNA chip covering ~10,000 rat genes from a total of approximately 30,000
- Measure gene expression changes across developmental stages and brain regions using DNA chip experiments
- Develop efficient unsupervised machine learning algorithms for DNA chip data analysis and expression cluster discovery
- Profile spatial expression patterns using in situ hybridization and validate computational findings
- Build a publicly accessible gene expression database with a web-based interface
Methodology
The project was organized across two annual phases:
Year 1
- DNA chip fabrication covering approximately 10,000 rat genes
- Analysis of gene expression variation across brain developmental stages and regions using DNA chips
- Development of DNA chip normalization algorithms (intensity ratio correction between Cy3 and Cy5 channels)
- Development of supervised clustering algorithms for DNA chip data
- Dependency analysis among gene expression patterns using Bayesian networks to model regulatory relationships
- Development of efficient learning algorithms leveraging neural network classifiers
- Design of a gene expression database schema
Year 2
- Analysis of gene expression changes in animal model systems
- Kernel-based learning algorithms for gene expression analysis using neural classifiers
- Temporal gene expression pattern modeling using DNA chip data
- Efficient analysis of in situ hybridization profile data
- Construction of a web-accessible gene expression database and user interface
Research Team
Principal Investigator
- Prof. Byoung-Tak Zhang (Seoul National University)
Researchers
- Sirk June Augh
- Kyu-Baek Hwang
- Seung-Woo Chung
- Dong-Min Kim
Contact
- Kyu-Baek Hwang — kbhwang@bi.snu.ac.kr
Collaboration
This project was conducted jointly with:
- School of Medicine, Hanyang University
- School of Medicine, Korea University
The collaboration bridged computational machine learning with experimental neurobiology, enabling both wet-lab DNA chip experiments and rigorous computational analysis of the resulting datasets.
Publications
- Kyu-Baek Hwang, Jae Won Lee, Seung-Woo Chung, and Byoung-Tak Zhang. Construction of large-scale Bayesian networks by local to global search. PRICAI 2002.