Toward a Multilevel Representation of Protein Molecules: Comparative Approaches to the Aggregation/Folding Propensity Problem

This paper builds upon the fundamental paper by Niwa et al. [33] that provides the unique possibility to analyze the relative aggregation/folding propensity of the elements of the entire Escherichia coli (E. coli) proteome in a cell-free standardized microenvironment. The hardness of the problem comes from the superposition between the driving forces of intraand inter-molecule interactions and it is mirrored by the evidences of shift from folding to aggregation phenotypes by single-point mutations [10]. Here in this paper we apply different state-of-the-art classification methods coming from the field of structural pattern recognition, with the aim to compare different representations of the same proteins of the Niwa et al. data base, going from pure sequence to chemico-physical labeled (contact) graphs. By this comparison, we are able to identify some interesting general properties of protein universe, going from the confirming of a threshold size around 250 residues (discriminating “easily foldable” from “difficultly foldable” molecules consistent with other independent data on protein domains architecture) to the relevance of contact graphs eigenvalue ordering for folding behavior discrimination and characterization of the E. coli data. The soundness of the experimental results presented in this paper is proved by the statistically relevant relationships discovered among the chemico-physical description of proteins and the developed cost matrix of substitution used in the various discrimination systems.