Alternative gene form discovery and candidate gene selection from gene indexing projects

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Gene indexing organizes expressed sequence tags (ESTs) and full-length transcripts into gene families.
• Accurate gene indexing is crucial for gene expression studies and discovering novel genes from ESTs.
• Existing methods can be affected by artifacts like chimerism, impacting index integrity.

Purpose of the Study:

• To enhance gene indexing by partitioning index classes into subclasses based on sequence dissimilarity.
• To improve the accuracy and utility of gene indices for downstream analyses.
• To demonstrate the method's ability to ameliorate artifacts and increase the sensitivity of gene expression studies.

Main Methods:

• Developed a subpartitioning technique to divide gene index classes into subclasses based on sequence diversity.
• Applied the subpartitioning method to the UniGene gene indexing project.
• Evaluated the impact of subpartitioning on index integrity, assembly quality, and gene expression analysis sensitivity.

Main Results:

• The subpartitioning method effectively ameliorated artifacts such as chimerism, improving index integrity.
• Treatment led to a marked increase in information quality and abundance in the UniGene dataset.
• Discovered new levels of information regarding differential expression of alternate gene forms, including regulated alternative splicing and polymorphisms.

Conclusions:

• Subpartitioning based on sequence dissimilarity significantly enhances gene indexing quality and information yield.
• This approach improves the detection of gene isoforms, polymorphisms, and tissue-specific gene regulation.
• The method offers a powerful tool for advancing gene expression studies and discovering novel biological insights.