CancerGram: An Effective Classifier for Differentiating Anticancer from Antimicrobial Peptides

Abstract

Antimicrobial peptides (AMPs) constitute a diverse group of bioactive molecules that provide multicellular organisms with protection against microorganisms, and microorganisms with weaponry for competition. Some AMPs can target cancer cells; thus, they are called anticancer peptides (ACPs). Due to their small size, positive charge, hydrophobicity and amphipathicity, AMPs and ACPs interact with negatively charged components of biological membranes. AMPs preferentially permeabilize microbial membranes, but ACPs additionally target mitochondrial and plasma membranes of cancer cells. The preference towards mitochondrial membranes is explained by their membrane potential, membrane composition resulting from α-proteobacterial origin and the fact that mitochondrial targeting signals could have evolved from AMPs. Taking into account the therapeutic potential of ACPs and millions of deaths due to cancer annually, it is of vital importance to find new cationic peptides that selectively destroy cancer cells. Therefore, to reduce the costs of experimental research, we have created a robust computational tool, CancerGram, that uses n-grams and random forests for predicting ACPs. Compared to other ACP classifiers, CancerGram is the first three-class model that effectively classifies peptides into: ACPs, AMPs and non-ACPs/non-AMPs, with AU1U amounting to 0.89 and a Kappa statistic of 0.65. CancerGram is available as a web server and R package on GitHub.

Keywords: anticancer peptide (ACP); anticancer peptides; antimicrobial peptide (AMP); antimicrobial peptides; host defense peptides; prediction; random forest.

Figure 1

Schematic representation of n-gram extraction (A) and decision-making procedure in CancerGram (B). The training data sets include ACP (shaded in red), AMP (shaded in yellow) and non-ACP/non-AMP sequences (the negative data set, shaded in blue). Each peptide from the training data sets was divided into subsequences of 5 amino acids (5-mers). For each 5-mer, we extracted continuous and discontinuous n-grams of size ranging from 1 to 3, and exemplary n-grams are presented in boxes shaded in colors respective to the data sets. The informative n-grams for CancerGram training were selected by Quick Permutation Test for all combinations of the data sets, and they are shaded in: (i) red-yellow for the ACP/AMP data set, (ii) red-blue for the ACP/Negative data set, and (iii) yellow-blue for the AMP/Negative data set (A). To make a prediction, CancerGram first divides a peptide into 5-mers and then, for each 5-mer, makes a prediction if it is an ACP, AMP or non-ACP/non-AMP (the first model). To scale the prediction from 5-mers to the level of a peptide, numerous statistics are calculated, and on their basis, CancerGram makes the final prediction (the second model) (B).

Figure 2

Distribution of the hydropathy index and net charge for anticancer peptides (ACPs), antimicrobial peptides (AMPs) and non-ACP/non-AMP sequences (Negative).

Figure 3

Amino acid composition of ACPs, AMPs and non-ACP/non-AMP sequences (Negative).

Figure 4

Results of fivefold cross-validation for the peptide layer of the model. Each dot corresponds to a single fold.