In this work, we study the correlation between codon usage and the network features of 11 the PPI in bacteria genomes. We want to extend the information by Dilucca et al. (2015) about 12 E.Coli's genome for a set of other 34 bacteria. We use PCA techniques in the space of codon bias 13 indices (compAI, compAI_w, tAI, NC) and GC content to show that genes with similar patterns of 14 codon usage feature have a significantly higher probability that their encoded proteins interact 15 within the PPI. And vice-versa, we show that interacting in the PPI have a coherent codon usage.
16This work could allow for future investigations into the possible effects that codon bias signal can 17 have on the topology of protein interaction network and, as such, to improve existing bioinformatics 18 methods for predicting protein interactions.
19Keywords: codon usage bias, bacteria, protein-protein interaction networks, interactomes 20 21
24As it is well-known, the genetic code is degenerate: synonymous codons at the genetic level encode 25 for the same amino acid in the translated protein. Moreover, although synonymous codons are 26 indistinguishable in the primary structure of a protein, they are not used randomly, but with 27 different frequencies which may vary across different species, in various parts of one specific 28 genome, and even within different regions of the same gene This phenomenon, known as Codon 29 Usage Bias (CUB), is well-established in the literature (see, for example, [3-5]). Nevertheless, while 30 remarkable observations have emerged, a general understanding of the biology of CUB still lacks,
31because of its complex phenomenology [7]. Although CUB does not alter the amino acid sequences 32 of proteins, it is involved in many important cellular processes, including differential gene 33 expression [7], translation efficiency and accuracy [8,9], dynamics of the ribosome, and 34 co-translational folding of the proteins [10,6].
35With the present study, we intend to share basic observations of sufficient generality on the 36 co-evolution of CUB and the degree connectivity of bacterial interactomes. The systematic analysis 37 of PPIs is crucial to understand the patterns of chemical reactions within cells, and the role played by 38 proteins in regulative processes. Moreover, on the applicative side, the comparison of the 39 interactomes from different species is fundamental to understand disease-related processes that 40 engage more than one species, such as host-pathogen relationships, to identify clinically relevant 41 host-pathogen PPI, and consequently to develop future therapeutic applications [1]. In previous 42 work, Dilucca et al. showed that translational selection in E. Coli (taken as a case of study) 129 variance of codon bias over the genomes. Thus, we focus on the plane defined by these two vectors.
130We calculate this plane for each bacterial genome (see Figure 2 in Supplementary Materials), 131 reporting the associated variances, eigenvalues and eigenvectors in Supplementary Materials. We 132 t...