Pattern recognition in nucleic acid sequences. I. A general method for finding local homologies and symmetries

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Identifying local similarities in biological sequences is crucial for understanding genome organization and function.
• Existing algorithms may not efficiently detect all types of local homologies and their statistical significance.

Purpose of the Study:

• To introduce a generalized algorithm for detecting local sequence similarities within longer sequences.
• To develop a method for statistically classifying these local homologies based on chance probability.
• To apply the algorithm for finding repeats, inverted repeats, and dyad symmetries in nucleic acid sequences.

Main Methods:

• Generalization of the Needleman-Wunsch-Sellers algorithm.
• Calculation of the probability of chance occurrence for local sequence resemblance.
• Statistical classification of local homologies.
• Analysis of random and biological nucleic acid sequences, including fourteen complete genomes.

Main Results:

• The algorithm successfully identifies local subsequences with significant resemblance.
• Statistical significance is accurately calculated, allowing for robust classification of homologies.
• The method effectively detects repeats, inverted repeats, and dyad symmetries.
• Application to complete genomes reveals patterns of dyad symmetry.

Conclusions:

• The developed algorithm provides a powerful tool for discovering statistically significant local homologies in biological sequences.
• It enhances the ability to identify various repetitive elements and symmetries within genomes.
• This approach has broad implications for sequence analysis and comparative genomics.