We get ready database out of fragments regarding a collection of low-redundant healthy protein domains
All of our sort of fragment selection and frequency integration gets sensible efficiency from secondary design anticipate checked-out towards 56 CASP5 needs (average SOV rating 0
We examined the consequences of various issues within the representing and you can combining evolutionary and you can structural guidance having local proteins architectural forecast considering fragment choices. For each and every fragment, evolutionary info is produced from homologous sequences and you will represented because projected productive counts and you can frequencies out-of amino acids (evolutionary frequencies) at each standing. Position-specific amino acid choice titled architectural wavelengths are based on statistical analysis off distinct local structural environment in the databases formations. All of our method for regional construction forecast is dependant on ranking and seeking database fragments that will be most like a goal fragment.
