Fast multiple alignment of ungapped DNA sequences using information theory and a relaxation method.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Identifying DNA binding sites for regulatory proteins like OxyR and Fis can be challenging due to dispersed sequence conservation.
• Traditional alignment methods struggle with identifying functionally relevant sites in such cases.

Purpose of the Study:

• To develop and present an information theory-based multiple sequence alignment method (Malign) for improved identification of DNA binding sites.
• To establish a robust and unambiguous criterion for evaluating the quality of sequence alignments.

Main Methods:

• Utilized an information theory-based multiple alignment algorithm (Malign) to analyze DNA binding sequences of OxyR and Fis proteins.
• Employed information content as a global quality criterion for alignment, using look-up tables for computational efficiency.
• Applied a hill-climbing algorithm to navigate the vast combinatorial space of possible alignments, starting from random selections.

Main Results:

• The Malign method provides an unambiguous way to select the best alignment using absolute units (bits) without arbitrary constants.
• The algorithm's speed allows for multiple starting points and classification of solutions, indicating convergence by a single, well-populated high-information content class.
• Distinct solution classes for Fis protein suggest self-similar features within its DNA binding sites.

Conclusions:

• The Malign algorithm offers a powerful and objective approach for multiple sequence alignment, particularly for identifying difficult-to-discern DNA binding sites.
• The method's reliance on information content provides a reliable metric for assessing alignment quality and biological relevance.
• Analysis of Fis protein binding sites reveals potential self-similar structural or functional characteristics.