Similarity Studies of Corona Viruses through Chaos Game Representation

In silico design of Ebola virus Glycoprotein antigenic peptides as vaccine candidates

Ebola virus (EBOV) is a filovirus that causes severe hemorrhagic fever and has a fatality rate between 50 and 90%. The vaccines were developed against the Ebola Zaire species; therefore, it is necessary to develop vaccines against other species to control future outbreaks. The objective of this work was to obtain vaccine candidate peptides against different EBOV species through the use of bioinformatics programs and servers that allow glycoprotein (GP) to be analyzed. GP sequences of various EBOV species that did not present gaps or unspecified amino acids or that were repeated (same year, region and laboratory) were downloaded from the NCBI database. A consensus sequence was generated and used to determine vaccine candidate peptides, which were evaluated, through a combination of servers and molecular dynamics, for their ability to interact with B and T lymphocytes, toxicity, allergenicity, solvent exposure, glycosylation, antigenicity, and presence in mature GP. Five vaccine candidate peptides were identified, of which PEP4 had the best characteristics evaluated in this study. PEP4 may be a potential candidate for the development of an EBOV vaccine.

CGRclust: Chaos Game Representation for twin contrastive clustering of unlabelled DNA sequences

BACKGROUND

Traditional supervised learning methods applied to DNA sequence taxonomic classification rely on the labor-intensive and time-consuming step of labelling the primary DNA sequences. Additionally, standard DNA classification/clustering methods involve time-intensive multiple sequence alignments, which impacts their applicability to large genomic datasets or distantly related organisms. These limitations indicate a need for robust, efficient, and scalable unsupervised DNA sequence clustering methods that do not depend on sequence labels or alignment.

RESULTS

This study proposes CGRclust, a novel combination of unsupervised twin contrastive clustering of Chaos Game Representations (CGR) of DNA sequences, with convolutional neural networks (CNNs). To the best of our knowledge, CGRclust is the first method to use unsupervised learning for image classification (herein applied to two-dimensional CGR images) for clustering datasets of DNA sequences. CGRclust overcomes the limitations of traditional sequence classification methods by leveraging unsupervised twin contrastive learning to detect distinctive sequence patterns, without requiring DNA sequence alignment or biological/taxonomic labels. CGRclust accurately clustered twenty-five diverse datasets, with sequence lengths ranging from 664 bp to 100 kbp, including mitochondrial genomes of fish, fungi, and protists, as well as viral whole genome assemblies and synthetic DNA sequences. Compared with three recent clustering methods for DNA sequences (DeLUCS, iDeLUCS, and MeShClust v3.0.), CGRclust is the only method that surpasses 81.70% accuracy across all four taxonomic levels tested for mitochondrial DNA genomes of fish. Moreover, CGRclust also consistently demonstrates superior performance across all the viral genomic datasets. The high clustering accuracy of CGRclust on these twenty-five datasets, which vary significantly in terms of sequence length, number of genomes, number of clusters, and level of taxonomy, demonstrates its robustness, scalability, and versatility.

CONCLUSION

CGRclust is a novel, scalable, alignment-free DNA sequence clustering method that uses CGR images of DNA sequences and CNNs for twin contrastive clustering of unlabelled primary DNA sequences, achieving superior or comparable accuracy and performance over current approaches. CGRclust demonstrated enhanced reliability, by consistently achieving over 80% accuracy in more than 90% of the datasets analyzed. In particular, CGRclust performed especially well in clustering viral DNA datasets, where it consistently outperformed all competing methods.

Exploring geometry of genome space via Grassmann manifolds

It is important to understand the geometry of genome space in biology. After transforming genome sequences into frequency matrices of the chaos game representation (FCGR), we regard a genome sequence as a point in a suitable Grassmann manifold by analyzing the column space of the corresponding FCGR. To assess the sequence similarity, we employ the generalized Grassmannian distance, an intrinsic geometric distance that differs from the traditional Euclidean distance used in the classical k-mer frequency-based methods. With this method, we constructed phylogenetic trees for various genome datasets, including influenza A virus hemagglutinin gene, Orthocoronavirinae genome, and SARS-CoV-2 complete genome sequences. Our comparative analysis with multiple sequence alignment and alignment-free methods for large-scale sequences revealed that our method, which employs the subspace distance between the column spaces of different FCGRs (FCGR-SD), outperformed its competitors in terms of both speed and accuracy. In addition, we used low-dimensional visualization of the SARS-CoV-2 genome sequences and spike protein nucleotide sequences with our methods, resulting in some intriguing findings. We not only propose a novel and efficient algorithm for comparing genome sequences but also demonstrate that genome data have some intrinsic manifold structures, providing a new geometric perspective for molecular biology studies.

CGRWDL: alignment-free phylogeny reconstruction method for viruses based on chaos game representation weighted by dynamical language model

Traditional alignment-based methods meet serious challenges in genome sequence comparison and phylogeny reconstruction due to their high computational complexity. Here, we propose a new alignment-free method to analyze the phylogenetic relationships (classification) among species. In our method, the dynamical language (DL) model and the chaos game representation (CGR) method are used to characterize the frequency information and the context information of k-mers in a sequence, respectively. Then for each DNA sequence or protein sequence in a dataset, our method converts the sequence into a feature vector that represents the sequence information based on CGR weighted by the DL model to infer phylogenetic relationships. We name our method CGRWDL. Its performance was tested on both DNA and protein sequences of 8 datasets of viruses to construct the phylogenetic trees. We compared the Robinson-Foulds (RF) distance between the phylogenetic tree constructed by CGRWDL and the reference tree by other advanced methods for each dataset. The results show that the phylogenetic trees constructed by CGRWDL can accurately classify the viruses, and the RF scores between the trees and the reference trees are smaller than that with other methods.

Polarization- and Chaos-Game-Based Fingerprinting of Molecular Targets of Listeria Monocytogenes Vaccine and Fully Virulent Strains

Two approaches to the synthesis of 2D binary identifiers ("fingerprints") of DNA-associated symbol sequences are considered in this paper. One of these approaches is based on the simulation of polarization-dependent diffraction patterns formed by reading the modeled DNA-associated 2D phase-modulating structures with a coherent light beam. In this case, 2D binarized distributions of close-to-circular extreme polarization states are applied as fingerprints of analyzed nucleotide sequences. The second approach is based on the transformation of the DNA-associated chaos game representation (CGR) maps into finite-dimensional binary matrices. In both cases, the differences between the structures of the analyzed and reference symbol sequences are quantified by calculating the correlation coefficient of the synthesized binary matrices. A comparison of the approaches under consideration is carried out using symbol sequences corresponding to nucleotide sequences of the hly gene from the vaccine and wild-type strains of Listeria monocytogenes as the analyzed objects. These strains differ in terms of the number of substituted nucleotides in relation to the vaccine strain selected as a reference. The results of the performed analysis allow us to conclude that the identification of structural differences in the DNA-associated symbolic sequences is significantly more efficient when using the binary distributions of close-to-circular extreme polarization states. The approach given can be applicable for genetic differentiation immunized from vaccinated animals (DIVA).

Quantification of the Diversity in Gene Structures Using the Principles of Polarization Mapping

Results of computational analysis and visualization of differences in gene structures using polarization coding are presented. A two-dimensional phase screen, where each element of which corresponds to a specific basic nucleotide (adenine, cytosine, guanine, or thymine), displays the analyzed nucleotide sequence. Readout of the screen with a coherent beam characterized by a given polarization state forms a diffracted light field with a local polarization structure that is unique for the analyzed nucleotide sequence. This unique structure is described by spatial distributions of local values of the Stokes vector components. Analysis of these distributions allows the comparison of nucleotide sequences for different strains of pathogenic microorganisms and frequency analysis of the sequences. The possibilities of this polarization-based technique are illustrated by the model data obtained from a comparative analysis of the spike protein gene sequences for three different model variants (Wuhan, Delta, and Omicron) of the SARS-CoV-2 virus. Various modifications of polarization encoding and analysis of gene structures and a possibility for instrumental implementation of the proposed method are discussed.

Accurate and fast clade assignment via deep learning and frequency chaos game representation

BACKGROUND

Since the beginning of the coronavirus disease 2019 pandemic, there has been an explosion of sequencing of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus, making it the most widely sequenced virus in the history. Several databases and tools have been created to keep track of genome sequences and variants of the virus; most notably, the GISAID platform hosts millions of complete genome sequences, and it is continuously expanding every day. A challenging task is the development of fast and accurate tools that are able to distinguish between the different SARS-CoV-2 variants and assign them to a clade.

RESULTS

In this article, we leverage the frequency chaos game representation (FCGR) and convolutional neural networks (CNNs) to develop an original method that learns how to classify genome sequences that we implement into CouGaR-g, a tool for the clade assignment problem on SARS-CoV-2 sequences. On a testing subset of the GISAID, CouGaR-g achieved an $96.29\%$ overall accuracy, while a similar tool, Covidex, obtained a $77,12\%$ overall accuracy. As far as we know, our method is the first using deep learning and FCGR for intraspecies classification. Furthermore, by using some feature importance methods, CouGaR-g allows to identify k-mers that match SARS-CoV-2 marker variants.

CONCLUSIONS

By combining FCGR and CNNs, we develop a method that achieves a better accuracy than Covidex (which is based on random forest) for clade assignment of SARS-CoV-2 genome sequences, also thanks to our training on a much larger dataset, with comparable running times. Our method implemented in CouGaR-g is able to detect k-mers that capture relevant biological information that distinguishes the clades, known as marker variants.

AVAILABILITY

The trained models can be tested online providing a FASTA file (with 1 or multiple sequences) at https://huggingface.co/spaces/BIASLab/sars-cov-2-classification-fcgr. CouGaR-g is also available at https://github.com/AlgoLab/CouGaR-g under the GPL.

Chaos game representation and its applications in bioinformatics

Chaos game representation (CGR), a milestone in graphical bioinformatics, has become a powerful tool regarding alignment-free sequence comparison and feature encoding for machine learning. The algorithm maps a sequence to 2-dimensional space, while an extension of the CGR, the so-called frequency matrix representation (FCGR), transforms sequences of different lengths into equal-sized images or matrices. The CGR is a generalized Markov chain and includes various properties, which allow a unique representation of a sequence. Therefore, it has a broad spectrum of applications in bioinformatics, such as sequence comparison and phylogenetic analysis and as an encoding of sequences for machine learning. This review introduces the construction of CGRs and FCGRs, their applications on DNA and proteins, and gives an overview of recent applications and progress in bioinformatics.

Prediction of antimicrobial resistance based on whole-genome sequencing and machine learning

MOTIVATION

Antimicrobial resistance (AMR) is one of the biggest global problems threatening human and animal health. Rapid and accurate AMR diagnostic methods are thus very urgently needed. However, traditional antimicrobial susceptibility testing (AST) is time-consuming, low throughput and viable only for cultivable bacteria. Machine learning methods may pave the way for automated AMR prediction based on genomic data of the bacteria. However, comparing different machine learning methods for the prediction of AMR based on different encodings and whole-genome sequencing data without previously known knowledge remains to be done.

RESULTS

In this study, we evaluated logistic regression (LR), support vector machine (SVM), random forest (RF) and convolutional neural network (CNN) for the prediction of AMR for the antibiotics ciprofloxacin, cefotaxime, ceftazidime and gentamicin. We could demonstrate that these models can effectively predict AMR with label encoding, one-hot encoding and frequency matrix chaos game representation (FCGR encoding) on whole-genome sequencing data. We trained these models on a large AMR dataset and evaluated them on an independent public dataset. Generally, RFs and CNNs perform better than LR and SVM with AUCs up to 0.96. Furthermore, we were able to identify mutations that are associated with AMR for each antibiotic.

AVAILABILITY AND IMPLEMENTATION

Source code in data preparation and model training are provided at GitHub website (https://github.com/YunxiaoRen/ML-iAMR).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.