Genome-wide analysis of single-nucleotide polymorphisms in human expressed sequences

Area of Science:

• Genomics
• Human Genetics
• Bioinformatics

Background:

• Single-nucleotide polymorphisms (SNPs) are crucial high-resolution markers for disease gene mapping.
• Expressed sequence tags (ESTs) offer a rich source for identifying genetic variations.

Purpose of the Study:

• To develop a robust method for detecting and validating a large set of SNPs in human coding regions.
• To create a public resource of SNPs to aid in disease gene discovery.

Main Methods:

• Utilized Bayesian inference to analyze human ESTs and identify candidate SNPs, distinguishing true polymorphisms from errors.
• Assessed various data quality metrics including chromatogram data, error rates, and cDNA library origin.
• Validated predicted SNPs through comparison with independently screened genes, HLA-A polymorphisms, and RFLP analysis.

Main Results:

• Identified 48,196 candidate SNPs primarily within coding regions of human genes.
• Achieved validation rates of 70%, 89%, and 71% across three independent verification methods.
• Detected significantly more true HLA-A SNPs compared to previous EST analyses.
• Found SNPs in a substantial proportion of known disease genes, including disease-causing mutations like the HbS sickle-cell mutation.

Conclusions:

• The developed statistical method effectively identifies and validates a large number of SNPs from EST data.
• The comprehensive SNP dataset represents a valuable public resource for accelerating the mapping of disease genes.
• This work enhances the utility of ESTs for discovering genetic variations relevant to human diseases.