Abstract
A pilot study was carried out to test the feasibility of using parallel cascade identification as a means for automatically distinguishing exon from intron DNA sequences. For the β T-cell receptor locus, only the first precisely determined intron and exon available were employed to form the input for training the parallel cascade models. The succeeding 25 introns and 28 exons with known boundaries were used as an evaluation set to discover the optimal values for four parameters that had to be preset before each model could be identified. The selected parallel cascade model attained a correct classification rate averaging about 82% on novel sequences from a test set and an unknown set used in a blind test.
Original language | English (US) |
---|---|
Pages (from-to) | 129-140 |
Number of pages | 12 |
Journal | Annals of Biomedical Engineering |
Volume | 30 |
Issue number | 1 |
DOIs | |
State | Published - 2002 |
Keywords
- DNA sequences
- Exons
- Identification
- Introns
- Nonlinear systems
ASJC Scopus subject areas
- Biomedical Engineering