<ul data-eligibleForWebStory="true"><li>Identification of critical residues of a protein is crucial for its function.</li><li>Three methods are presented to identify critical residues of a protein using Random Geometric Graphs.</li><li>A Random Geometric Graph is learned based on correlations between state variables of residues.</li><li>Cramer's V is used to compute the correlation between residue pairs.</li><li>Organic thresholding is introduced to learn the Random Geometric Graph.</li><li>Results are compared with existing thresholding techniques using nodal degree as a criticality parameter.</li><li>A criticality measure is developed by ranking differences in posterior probabilities between full and partial graphs.</li><li>Criticality is also studied based on the variation of nodal degrees as the protein evolves.</li><li>The study compares results obtained from the three criticality parameters with experimentally verified critical residues.</li></ul>

Identifying critical residues of a protein using meaningfully-thresholded Random Geometric Graphs

Discover more