Binary Matrix Method to Enumerate, Hierarchically Order, and Structurally Classify Peptide Aggregation

Protein aggregation is a common and
complex phenomenon in biological
processes, yet a robust analysis of this aggregation process remains
elusive. The commonly used methods such as center-of-mass to center-of-mass
(COM–COM) distance, the radius of gyration (Rg), hydrogen bonding (HB), and solvent accessible surface
area do not quantify the aggregation accurately. Herein, a new and
robust method that uses an aggregation matrix (AM) approach to investigate
peptide aggregation in a MD simulation trajectory is presented. An nxn two-dimensional AM is created by using the interpeptide Cα–Cα cutoff distances, which are binarily encoded (0 or 1). These aggregation
matrices are analyzed to enumerate, hierarchically order, and structurally
classify the aggregates. Comparison of the present AM method suggests
that it is superior to the HB method since it can incorporate nonspecific
interactions and the Rg and COM–COM
methods since the cutoff distance is independent of the length of
the peptide. More importantly, the present method can structurally
classify the peptide aggregates, which the conventional Rg, COM–COM, and HB methods fail to do. The unique
selling point of this method is its ability to structurally classify
peptide aggregates using two-dimensional matrices.