LogOddsLogo

LogOddsLogo computes statistically grounded sequence logos from DNA and protein multiple sequence alignments to identify biologically significant alignment columns.

Key Features:

• Log-Odds Scoring: Computes per-observation multiple-alignment log-odds scores as the information content at each alignment position, assigning positive or negative values to indicate likelihood of biological relatedness versus random chance.
• Statistical Measures: Implements normalized maximum likelihood and Bayesian measures to assess and normalize the significance of column information.
• Improved Discrimination in Protein Alignments: Enhances discrimination of functionally important positions in protein alignments, demonstrated on High Mobility Group B (HMGB) box proteins and enzyme datasets.

Scientific Applications:

• Sequence alignment interpretation: Provides statistically rigorous sequence logos to aid accurate interpretation of DNA and protein multiple sequence alignments.
• Protein function and evolution: Identifies biologically relevant alignment columns to support studies of protein function and evolutionary conservation.

Methodology:

Per-position information is computed as per-observation multiple-alignment log-odds scores and assessed using normalized maximum likelihood and Bayesian measures.