Читать книгу Systematics and the Exploration of Life - Группа авторов - Страница 29

The methodology presented allows the identification of disturbed regions in protein structures by taking into account biases due to experimental variations and protein flexibility. Now that we know that mutations do indeed disrupt the main chain and that these disruptions are measurable with current techniques, it would be interesting to model them, especially to improve the predictions of ΔΔG, for which the carbon chain is rigid.

Two models exist for the accommodation of the main chain under the effect of amino acid substitution. The first (Davis et al. 2006) is derived from the observation of alternative atomic positions in ultra-high resolution crystallographic structures. It has been successfully used to improve Rosetta’s calculation of ΔΔG (Lauck et al. 2010). The second (Bordner and Abagyan 2004) was constructed from data collected on 2,141 pairs of protein structures, only differing by a single point mutation. This model also improved Gibbs’ prediction of free energy after a mutation. The selection method presented allows the identification of fragments where the main chain was more disrupted than expected. Using this database instead of the previous ones should improve the models.