Software for structure annotation has become a key aspect for any functional study, as structure is highly conserved and a vast amount of structural data is available. Implemented in software like DSSP, Cablam or Gesamt, algorithms employed for structure analysis are based either on sequence features, geometrical parameters or a combination of them. For the validation context, algorithms based on geometry descriptors have been developed, such as C-Alpha Based Low-resolution Annotation Method (CABLAM)[1]. ALEPH addresses two challenges:
ALEPH, which incorporates the previous algorithm Borges [2], has been developed to analyze small local folds. ALEPH is integrated in the ARCIMBOLDO software for ab initio phasing, especially in ARCIMBOLDO_SHREDDER [3] for identification of compact folds. The software uses ad hoc parameters called Characteristic Vectors (CV's) for retrieving tridimensional properties of the main chain. CV's exploit the fact that carbonyl oxygens are differently oriented for each secondary structure. Properties of the CV such as their modulus or direction, the angles between them or the distances that separates them, allow to identify different secondary structures and describe their environment. Recent developments, including an increase in the number of parameters used for main chain description and the addition of new mathematical algorithms has allowed not only to improve the precision in the annotation but also its flexibility.
Beyond secondary structure annotation, ALEPH can also optimally superpose very small structures, annotate and extract folds from a model or generate libraries of a given fold. A general description of the performance of ALEPH for validation of non ideal main chains, identification of a fold from an unknown solution and identification of compact structural folds in the frame of ARCIMBOLDO_SHREDDER will be illustrated.
References: