bioRxiv Science⌕ Search

bioRxiv · 10.1101/697243

Machine-learning classification suggests that many alphaproteobacterial prophages may instead be gene transfer agents

Abstract

Many of the sequenced bacterial and archaeal genomes encode regions of viral provenance. Yet, not all of these regions encode bona fide viruses. Gene transfer agents (GTAs) are thought to be former viruses that are now maintained in genomes of some bacteria and archaea and are hypothesized to enable exchange of DNA within bacterial populations. In Alphaproteobacteria, genes homologous to the head-tail gene cluster that encodes structural components of the Rhodobacter capsulatus GTA (RcGTA) are found in many taxa, even if they are only distantly related to Rhodobacter capsulatus. Yet, in most genomes available in GenBank RcGTA-like genes have annotations of typical viral proteins, and therefore are not easily distinguished from their viral homologs without additional analyses. Here, we report a support vector machine classifier that quickly and accurately distinguishes RcGTA-like genes from their viral homologs by capturing the differences in the amino acid composition of the encoded proteins. Our open-source classifier is implemented in Python and can be used to scan homologs of the RcGTA genes in newly sequenced genomes. The classifier can also be trained to identify other types of GTAs, or even to detect other elements of viral ancestry. Using the classifier trained on a manually curated set of homologous viruses and GTAs, we detected RcGTA-like head-tail gene clusters in 57.5% of the 1,423 examined alphaproteobacterial genomes. We also demonstrated that more than half of the in silico prophage predictions are instead likely to be GTAs, suggesting that in many alphaproteobacterial genomes the RcGTA-like elements remain unrecognized.\n\nData depositionSequence alignments and phylogenetic trees are available in a FigShare repository at DOI 10.6084/m9.figshare.8796419. The Python source code of the described classifier and additional scripts used in the analyses are available via a GitHub repository at https://github.com/ecg-lab/GTA-Hunter-v1

Source connections

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Kogay, R., Neely, T. B., Birnbaum, D. P., Hankel, C. R., Shakya, M., Zhaxybayeva, O.. 2019-07-11. Machine-learning classification suggests that many alphaproteobacterial prophages may instead be gene transfer agents. https://doi.org/10.1101/697243

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

SPACE-Tag Enables High-Throughput Spatial Chromatin Profiling With CUT&Tag

Spatial epigenomics maps chromatin states within intact tissues, yet current methods remain limited by accessibility and scale. We developed SPACE-Tag, a spatial CUT&Tag framework that converts epigenetic marks-associated DNA into polyadenylated RNA through T7 based linear amplification, enabling direct integration with standard spatial transcriptomic workflows. Applied to the mouse brain, SPACETag resolved activating and repressive chromatin landscapes, recovered region-specific gene regulatory programs across anatomical domains, and showed strong correlation with existing bulk and single cell reference data. We generated an integrated multimodal atlas of the aging mouse brain, profiling five histone modifications and the transcriptome in parallel across the adult lifespan to establish the largest spatial multimodal dataset to date. This analysis uncovered sharp regional divergence with aging in the brain: fiber tracts showed widespread transcriptional downregulation, often without nearby chromatin state transitions, whereas the isocortex showed coordinated chromatin transcriptome changes, including an H3K4me1 peak module that increased with age, was enriched at super-enhancers and linked to upregulation of neuronal plasticity regulators. These findings highlight the power of multimodal and high throughput spatial genomics in uncovering the divergent regulatory logic that governs tissue-specific vulnerability and resilience in the aging brain.

genomics↗

Long-read DNA sequencing reveals global hypomethylation and focal hypermethylation in SDHx-deficient pheochromocytoma and paraganglioma

Pheochromocytomas and paragangliomas (PPGL) are highly heritable neuroendocrine tumors of the autonomic ganglia. Genomic DNA in most cancer cells is hypomethylated. However, DNA in PPGLs secondary to mutations in succinate dehydrogenase complex (SDHx) genes is reportedly hypermethylated due to succinate-mediated inhibition of DNA demethylases. We used nanopore long-read sequencing to characterize the methylome of tumor and white blood cell DNA from SDHx and non-SDHx PPGL patients, alongside normal adrenal medulla. We find that both SDHx and non-SDHx PPGLs are globally hypomethylated, particularly at centromeres. In normal adrenal medulla, CpG islands marked by H3K4 methylation are hypomethylated, whereas CpG islands lacking H3K4 methylation are hypermethylated. In SDHx-related PPGL, some H3K4me-marked CpG islands are hypermethylated, but most are not. We propose that DNA methylation in SDHx-related PPGL reflects the global hypomethylation generally observed in tumors, partly offset by succinate-mediated inhibition of TET DNA demethylases. Opposing effects of succinate may alter the balance between DNA methylation (indirect inhibition of the DNMT3A/B DNA methylases via inhibition of H3K4me demethylases) and DNA demethylation (inhibition of TET DNA demethylases) at active CpG islands.

genomics↗

High piRNA abundance is a hallmark of evolutionarily persistent piRNA-producing loci.

PIWI-interacting RNAs (piRNAs) silence transposons in the germline, and the loci that produce them diverge rapidly between species. Yet, some loci retain both their genomic location and their piRNA production across large evolutionary distances, and the basis for this conservation is unclear. To identify the features that distinguish these persistent loci, we compared piRNA production at orthologous loci in three closely related mouse species, contrasting loci that produce piRNAs in all three species with those that do so in only one. Transposon content and UTR features, often linked to piRNA production, did not distinguish the two groups. Transcription extending beyond gene ends (readthrough) was more frequent at conserved loci, but did not explain conservation once the amount of piRNAs produced was taken into account. Instead, we found that the conserved loci are those that produce the most piRNAs. The same association holds true in human and two other primate species. We speculate that most piRNAs expressed in testis are tolerated molecular noise, and that only the highly expressed minority are likely to be functional.

genomics↗