Amino Acid Sequence Autocorrelation Vectors and Ensembles of Bayesian‐Regularized Genetic Neural Networks for Prediction of Conformational Stability of Human Lysozyme Mutants.

Abstract

ChemInform is a weekly Abstracting Service, delivering concise information at a glance that was extracted from about 200 leading journals. To access a ChemInform Abstract, please click on HTML or PDF.

Proteometric study of ghrelin receptor function variations upon mutations using amino acid sequence autocorrelation vectors and genetic algorithm-based least square support vector machines

“…As weights for sequence residues they were used 48 physicochemical, energetic, and conformational amino acid/ residues properties (Table 1) selected by Gromiha et al [21] from the AAindex database [12] in a previous study concerning relationships between amino acid/residues properties and protein stability for a large set of proteins. These properties were recently used by us for generating human lysozymes AASA vectors for modeling conformational stability [11] and by Grommiha et al [22] for predicting with protein folding rates. In our work, spatial lag, l, was ranging from 1 to 15 with the aim of accessing to long-range interactions in the sequence due to tertiary structure arrangements.…”

“…By means of a GA-based QSAR study we aimed to identify relevant structural features influencing the activity of the ghrelin receptor. In this connection, the novel reported AASA vectors were used for protein structural information encoding [11]. Amino acid sequence of human ghrelin receptor (primary accession number Q92847) was obtained from the Swiss-Prot/ In white on black are the three Phe residues that were subjected to more elaborate mutagenesis.…”

“…Similarly, a novel bioinformatics approach named as proteochemometrics including the mapping of molecular recognition and not necessary needing knowledge of the three-dimensional structure of biomacromolecules has successfully applied to the study of protein-ligand interactions [9]. In this context, we recently introduced the amino acid sequence autocorrelation (AASA) vectors for modeling the conformational stability of human lysozyme mutants [11]. AASA vectors are weighted by 48 physicochemical, energetic, and conformational amino acid/ residues properties extracted from the AAindex amino acid database [12].…”

“…QSAR studies of proteins have been developed by extending conventional graph-theoretical representation of chemical structures to protein sequence and three-dimensional structure in combination with statistical methods for regression and/or classification analysis [10,11]. This approach has been applied to protein stability studies and protein classification.…”

Protein radial distribution function (P-RDF) and Bayesian-Regularized Genetic Neural Networks for modeling protein conformational stability: Chymotrypsin inhibitor 2 mutants

“…Artificial Neural Networks (ANNs) usually overcome methods limited to linear regression models like MRA or Partial Least Square [22][23][24][25][26][27][28]. In this connection, we recently extend the concept of structural autocorrelation vectors in molecules to protein sequences and ensembles of Bayesian-Regularized Genetics Neural Networks (BRGNN) successfully modeled conformational stability of human lysozyme [29] and gene V protein [30] mutants.…”