Dihedral angle database based on disulfide bonds

• Main Authors: Dayalan, S., Bevinakoppa, S.
• Format: Conference Proceeding
• Language: English
• Subjects: Amino acids; Australia; Bioinformatics; Bonding; Computer science; Information technology; Protein sequence; Spine; Statistical analysis
• DOI: 10.1109/BIBE.2004.1317376

Dihedral angle databases are protein databases that store all occurring dihedral angle (phi, psi) values of amino acids in proteins. None of the existing dihedral angle databases have classified their data based on an important bonding characteristic of proteins called disulfide bonds. In this paper, using statistical analysis, the need to classify values in a dihedral angle database based on disulfide bonds is shown. This paper discusses how our dihedral angle database (DAB) is classified into two; dihedral database (DABSS) having proteins with at least one disulfide bond and dihedral database (DABNSS) having proteins without any disulfide bonds. Statistical analysis is used to show that the dihedral values (and hence the structure) of sub-sequences of amino acids obtained from DABSS are significantly different from their corresponding values in DABNSS, thus justifying the need to classify data based on disulfide bonds. Using statistical analysis to show that the values in DAB are significantly different from the corresponding values in DABSS and DABNSS strengthens this justification. Using this analysis and by querying DAB, DABSS and DABNSS it is shown how the values obtained by a protein structure prediction program yields different results and how querying DAB gives mixed results as it contains dihedral values from proteins with and without disulfide bonds together.