A New Alignment-Free Approach to Compare Protein Secondary Structure Families through SSEs

Understanding the similarity and dissimilarity among protein secondary structural families is essential for interpreting their structural features and functional roles. This study presents a new alignment-free approach for comparing protein families based on their Secondary Structure Elements (SSEs), classified into helix (H), strand (E), and coil (C). By numerically representing these SSEs, a four-component descriptor is derived using the moment of inertia, enabling the computation of distance matrices through four distinct distance metrics: Manhattan, Standard, Euclidean, and Maximum. Phylogenetic trees are subsequently constructed using the Neighbor-Joining (NJ) method. The methodology is evaluated across six benchmark datasets of various length representing four protein superfamilies, sourced from the Structural Classification of Proteins (SCOP). Among the distance measures tested, the Manhattan metric consistently yielded the most reliable results in capturing the relationships between structural families. Comparative analysis with other metrics and existing approaches underscores the robustness and effectiveness of the proposed method. This work demonstrates that the introduced descriptor and comparative framework provide a reliable, quantitative means of assessing structural similarity among protein SSEs of varying length, offering a valuable alignment-free method for structural bioinformatics and protein classification.