Assessing Iterative Normal Modes on Representing Protein Conformational Transitions

How to capture biologically important conformational transitions from structures have been of key interest in computational biology. It is generally recognized that normal modes derived from structure-based potentials can well represent functional movements in lots of biomolecules. Recently normal mode analysis was successfully used in an iterative manner to explore protein conformational transition pathway, but the impact of these iterative normal modes on representing structural movements is not clear now. To reveal the characteristics of these modes, we generate conformational transition pathways by freely deforming protein conformations along selected low-frequency normal modes, and then compare these new modes. Through this simple calculation, we find that although the low-frequency normal modes obtained from intermediate conformations can capture currently structural movements toward target states in a wide range of structural areas, their ability decreases gradually when protein moves away from reference structure.