Fusing Spectral Data To Improve Protein Secondary Structure Analysis: Data Fusion

The determination of protein secondary structure has become an area of great significance as this knowledge is important for understanding relationships between protein structure and, more importantly, how the changes in structure affect function. Previous studies suggest that a complementary use of spectroscopic data from optical methods such as circular dichroism (CD), infrared (IR) and ultraviolet resonance Raman (UVRR) coupled with multivariate calibration techniques like multivariate curve resolution-alternating least squares (MCR-ALS) is the preferred route for real-time and accurate evaluation of protein secondary structure. This study presents a new strategy for the improvement of secondary structure determination of proteins by fusing CD and UVRR spectroscopic data. Also, a new method for determining the structural composition of each protein is employed, which is based on the relative abundance of the ((pm) dihedral angles of the peptide backbone as they correspond to each type of secondary structure.. Comparison of the predicted protein secondary structures from MCR-ALS analysis of CD, UVRR and fused data with definitions obtained from dihedral angles of the peptide backbone, yields lower overall root mean squared errors of calibration for helical, beta-sheet, poly-proline II type and total unfolded secondary structures with fused data.