I try to make a model with MODELLER software and I decide to use multiple template modeling method. Two template structure I choose for my work, have similar Identity percentage (42.54% and 42.27%), identical coverage, different scores (583 and 566) and different resolution (2.25 and 1.55 Angstrom). I`m confused which one is better for choosing main template: score or structure resolution.
At such similar scores, either template should work equally well. However, compare the structures of the potential template in a molecular graphics application to look for structural differences and differences in coverage (e.g. unresolved loops or flexible sequences being visible only because they are held in place by crystal contacts). Additional things to look at depend on the purpose of your models. e.g. if you plan to use your model for docking, a liganded template structure may be preferable to an unliganded on, especially if the accessibility or size of the ligand binding pocket differs between the two.
In addition, at this level of identity, you may in addition to Modeller, also want to try modelling with a threading algorithm, e.g. iTasser, which not only rely on how similar two sequences are using average amino acid similarities, but actually check how "happy" (energetically favourable) your sequence is in the structure of the template. Especially if you have man insertions/deletions between your target sequence and the sequence of your template, Modeller is crucially dependent on the correct placement of these gaps, which is not necessarily the on yielding the best score.
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling