Biology subjects

Tolstoy, I.

Publications and source records attributed to Tolstoy, I..

6 recordsLinked to original sources

bioRxiv2025

A novel approach to Caudoviricetes taxonomy utilising whole proteome structure-structure comparison

Viral proteins often evolve so rapidly that sequence similarity is lost even among conserved functional genes, complicating phylogenetic analysis and classification. This challenge is acute for Caudoviricetes, a highly diverse class of dsDNA tailed bacteriophages recently redefined by the International Committee on Taxonomy of Viruses (ICTV). To address the limitations of sequence-based taxonomy, we present a structure-based classification framework using whole-proteome structural clustering. From 4,082 exemplar genomes, we predicted 445,098 protein structures with ESMFold and clustered them using Foldseek. Genomes were encoded as binary profiles of structural fold presence, and phylogenetic distances were inferred and normalized using Relative Evolutionary Divergence (RED). This yielded a revised structural taxonomy comprising 159 orders, 267 families, 502 subfamilies, and 1,189 genera. We also introduce PhagePleats, a Python tool for classifying novel phage genomes based on structural similarity. Our approach highlights the utility of protein structure for resolving distant viral relationships.

microbiology↗

bioRxiv2024

Muscle-3D: scalable multiple protein structure alignment

Protein multiple alignment is an essential step in many bioinformatics analysis such as phylogenetic tree estimation, HMM construction and critical residue identification. Structure is conserved between distantly-related proteins where amino acid similarity is weak or undetectable, suggesting that structure-informed sequence alignments might offer advantages over alignments constructed from amino acid sequences alone. The advent of the AI folding era has unleashed millions of high-quality predicted structures, motivating the development and assessment of scalable multiple structure alignment (MStA) methods. Here, we describe Muscle-3D, a new MStA algorithm combining a rich sequence representation of structure context, the Reseek "mega-alphabet", with state-of-the art alignment techniques from Muscle5 including a posterior decoding pair-HMM, consistency transformation, iterative refinement and ensemble construction. We show that Muscle-3D readily scales to thousands of structures. Comparative validation on several benchmark datasets using different quality metrics shows Muscle-3D to be among the higher-scoring methods, but we find that algorithm rankings from different metrics disagree despite low P-values according to the Wilcoxon rank-sum test. We suggest that these conflicts arise from the inherently fuzzy nature of structural alignment, and argue that a universal standard of MStA accuracy is not possible in principle. We describe contact map profiles for visualizing variation in inter-residue distances, and introduce a novel measure of local conformation similarity, LDDT-muw. Muscle-3D software is available at https://github.com/rcedgar/muscle.

bioinformatics↗

bioRxiv2022

Type IV pili is a critical virulence factor in clinical isolates of Paenibacillus thiaminolyticus

Hydrocephalus, the leading indication for childhood neurosurgery worldwide, is particularly prevalent in low-and-middle-income countries (LMICs). Hydrocephalus preceded by an infection, or postinfectious hydrocephalus (PIH), accounts for up to 60% of hydrocephalus in LMICs. Since many children with hydrocephalus suffer poor long-term outcomes despite surgical intervention, prevention of hydrocephalus remains paramount. Our previous studies implicated a novel bacterial pathogen, Paenibacillus thiaminolyticus, as a contributor to PIH in Uganda. Here we report the isolation of three P. thiaminolyticus strains, Mbale, Mbale2, and Mbale3, from patients with PIH and the demonstration that the three clinical isolates exhibit virulence in mice while P. thiaminolyticus type strain, B-4156, does not. We constructed complete genome assemblies of the clinical isolates as well as the reference strain and performed comparative genomics and proteomics analyses to identify potential virulence factors. One candidate virulence factor is a cluster of genes carried on a mobile genetic element that encodes a type IV pilus and is present in all three PIH patient strains but absent in the type strain. Proteomic and transcriptomic data confirmed the expression of this cluster of genes in the Mbale strain, while CRISPR-mediated deletion of the gene cluster substantially reduced the virulence of this strain. Our comparative proteogenomic analysis also identified various antibiotic resistance loci in the virulent strains. These results provide insight into the mechanism of virulence of Paenibacillus thiaminolyticus and suggest avenues for the diagnosis and treatment of this novel bacterial pathogen. Author SummaryPostinfectious hydrocephalus (PIH), a devastating sequela of neonatal infection, is associated with increased childhood mortality and morbidity. Paenibacillus thiaminolyticus was recently identified as the dominant organism highly associated with PIH in an African cohort. Our whole-genome sequencing, RNA sequencing and proteomics of three clinical isolates and a type strain in combination with CRISPR editing has revealed the type IV pili (T4P), encoded in a mobile genetic element, as a critical virulence factor for P. thiaminolyticus infection. Given the widespread presence of T4P in pathogens, the presence of T4P operon could serve as an important diagnostic and therapeutic target in P. thiaminolyticus and related bacteria.

microbiology↗

bioRxiv2020

Unique genomic features of crAss-like phages, the dominant component of the human gut virome

CrAssphage is the most abundant virus identified in the human gut virome and the founding member of a large group of bacteriophages that infect bacteria of the phylum Bacteroidetes and have been discovered by metagenomics of both animal-associated and environmental habitats. By analysis of circular contigs from human gut microbiomes, we identified nearly 600 genomes of crAss-like phages. Phylogenetic analysis of conserved genes demonstrates the monophyly of crAss-like phages, which can be expected to become a new order of viruses, and of 5 distinct branches, likely, families within that order. Two of these putative families have not been identified previously. The phages in one of these groups have large genomes (145-192 kilobases) and contain an unprecedented high density of self-splicing introns and inteins. Many crAss-like phages encode suppressor tRNAs that enable readthrough of UGA or UAG stop-codons, mostly, in late phage genes, which could represent a distinct anti-defense strategy. Another putative anti-defense mechanism that might target an unknown defense system in Bacteroidetes inhibiting phage DNA replication involves multiple switches of the phage DNA polymerase type between A and B families. Thus, comparative genomic analysis of the expanded assemblage of crAss-like phages reveals several unusual features of genome architecture and expression as well as phage biology that were not apparent from the previous crAssphage analyses.

microbiology↗

bioRxiv2019

Using the NCBI AMRFinder Tool to Determine Antimicrobial Resistance Genotype-Phenotype Correlations Within a Collection of NARMS Isolates

Antimicrobial resistance (AMR) is a major public health problem that requires publicly available tools for rapid analysis. To identify acquired AMR genes in whole genome sequences, the National Center for Biotechnology Information (NCBI) has produced a high-quality, curated, AMR gene reference database consisting of up-to-date protein and gene nomenclature, a set of hidden Markov models (HMMs), and a curated protein family hierarchy. Currently, the Bacterial Antimicrobial Resistance Reference Gene Database contains 4,579 antimicrobial resistance gene proteins and more than 560 HMMs. Here, we describe AMRFinder, a tool that uses this reference dataset to identify AMR genes. To assess the predictive ability of AMRFinder, we measured the consistency between predicted AMR genotypes from AMRFinder against resistance phenotypes of 6,242 isolates from the National Antimicrobial Resistance Monitoring System (NARMS). This included 5,425 Salmonella enterica, 770 Campylobacter spp., and 47 Escherichia coli phenotypically tested against various antimicrobial agents. Of 87,679 susceptibility tests performed, 98.4% were consistent with predictions. To assess the accuracy of AMRFinder, we compared its gene symbol output with that of a 2017 version of ResFinder, another publicly available resistance gene database. Most gene calls were identical, but there were 1,229 gene symbol differences between them, with differences due to both algorithmic differences and database composition. AMRFinder missed 16 loci that Resfinder found, while Resfinder missed 1,147 loci AMRFinder identified. Two missing drug classes from the 2017 version of ResFinder contributed 81% of missed loci. Based on these results, AMRFinder appears to be a highly accurate AMR gene detection system. ImportanceAntimicrobial resistance is a major public health problem. Traditionally, antimicrobial resistance has been identified using phenotypic assays. With the advent of genome sequencing, we now can identify resistance genes and deduce if an isolate could be resistant to antibiotics. We describe a database of 4,579 acquired antimicrobial resistance genes, the largest publicly available, and a software tool to identify genes in bacterial genomes, AMRFinder. Unlike other tools, AMRFinder uses a gene hierarchy to prevent overpredicting what the correct gene call should be, enabling more accurate assessment. To assess these resources, we determined the resistance gene content of over 6,200 bacterial isolates from the National Antimicrobial Resistance Monitoring System that have been assayed using traditional methods and that also have had their genomes sequenced. We also compared our gene assessments to those of a popularly used tool. We found that AMRFinder has a high overall consistency between genotypes and phenotypes.

microbiology↗

bioRxiv2017

Genomic, proteomic, and phylogenetic analysis of spounaviruses indicates paraphyly of the order Caudovirales

It is almost a cliche that tailed bacteriophages of the order Caudovirales are the most abundant and diverse viruses in the world. Yet, their taxonomy still consists of a single order with just three families: Myoviridae, Siphoviridae, and Podoviridae. Thousands of newly discovered phage genomes have recently challenged this morphology-based classification, revealing that tailed bacteriophages are genomically even more diverse than once thought. Here, we evaluate a range of methods for bacteriophage taxonomy by using a particularly challenging group as an example, the Bacillus phage SPO1-related viruses of the myovirid subfamily Spounavirinae. Exhaustive phylogenetic and phylogenomic analyses indicate that the spounavirins are consistent with the taxonomic rank of family and should be divided into at least five subfamilies. This work is a case study for virus genomic taxonomy and the first step in an impending massive reorganization of the tailed bacteriophage taxonomy.

microbiology↗