Related Experiment Videos
String analysis and energy minimization in the partition of DNA sequences
Journal of Molecular Biology
|June 5, 1989
Summary
This study shows that local patterns in DNA sequences can predict their behavior, even when based on global physical properties. Expert systems effectively classified sequences, demonstrating the link between physical constraints and observable patterns.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Understanding biological sequences often involves analyzing literal sequence motifs (strings).
- Alternatively, sequence recognition relies on complex physical interactions and global properties.
Purpose of the Study:
- To evaluate the potential of using local pattern analysis to predict sequence behavior based on global physical properties.
- To investigate the relationship between physical constraints, evolutionary processes, and observable sequence patterns.
Main Methods:
- Classified DNA sequences as positive or negative based on a global physical property (single melted domain).
- Generated positive biological sequences via computer simulation of evolutionary divergence.
- Applied pattern analysis and constructed expert systems to discriminate between positive and negative sequences.
Main Results:
- Expert systems achieved high accuracy (79-99%) in classifying sequences.
- Errors on negative sequences were less than 2%.
- Global physical constraints can generate local patterns sufficient for prediction, but large datasets are needed.
Conclusions:
- Local sequence patterns, driven by global physical constraints and evolutionary processes, can reliably predict sequence behavior.
- The effectiveness of string analysis depends on the specific evolutionary trajectory and sequence properties.
- The developed test-case is valuable for studying sequence evolution and pattern analysis techniques.