STALLION: a stacking-based ensemble learning framework for prokaryotic lysine acetylation site prediction

Protein post-translational modification (PTM) is an important regulatory mechanism that plays a key role in both normal and disease states. Acetylation on lysine residues is one of the most potent PTMs owing to its critical role in cellular metabolism and regulatory processes. Identifying protein lysine acetylation (Kace) sites is a challenging task in bioinformatics. To date, several machine learning-based methods for the in silico identification of Kace sites have been developed. Of those, a few are prokaryotic species-specific. Despite their attractive advantages and performances, these methods have certain limitations. Therefore, this study proposes a novel predictor STALLION (STacking-based Predictor for ProkAryotic Lysine AcetyLatION), containing six prokaryotic species-specific models to identify Kace sites accurately. To extract crucial patterns around Kace sites, we employed 11 different encodings representing three different characteristics. Subsequently, a systematic and rigorous feature selection approach was employed to identify the optimal feature set independently for five tree-based ensemble algorithms and built their respective baseline model for each species. Finally, the predicted values from baseline models were utilized and trained with an appropriate classifier using the stacking strategy to develop STALLION. Comparative benchmarking experiments showed that STALLION significantly outperformed existing predictor on independent tests. To expedite direct accessibility to the STALLION models, a user-friendly online predictor was implemented, which is available at: http://thegleelab.org/STALLION.

Identifying small ORFs in prokaryotes with ribosome profiling

Small proteins encoded by ORFs shorter than 50 codons (sORFs) are often overlooked by annotation engines and are difficult to characterize using traditional biochemical techniques. Ribosome profiling has tremendous potential to empirically improve the annotations of prokaryotic genomes. Recent improvements in ribosome profiling methods for bacterial model organisms have revealed many new sORFs in well-characterized genomes. Antibiotics that trap ribosomes just after initiation have played a key role in these developments by allowing unambiguous identification of the start codons (and hence the reading frame) for novel ORFs. Here we describe these new methods and highlight critical controls and considerations for adapting ribosome profiling to different prokaryotic species.

Metagenome-Assembled Genome Sequences of Novel Prokaryotic Species from the Mercury-Contaminated East Fork Poplar Creek, Oak Ridge, Tennessee, USA

We sequenced two metagenomes of sediments from the East Fork Poplar Creek in the Oak Ridge Reservation (Oak Ridge, TN), a natural stream that has been contaminated with Hg from upstream sources, and we reconstructed 28 metagenome-assembled genomes of novel prokaryotic species.

Human reference gut microbiome comprising 5,414 prokaryotic species, including newly assembled genomes from under-represented Asian metagenomes

Metagenome sampling bias for geographical location and lifestyle is partially responsible for the incomplete catalog of reference genomes of gut microbial species. Here, we present a substantially expanded microbiome catalog, the Human Reference Gut Microbiome (HRGM). Incorporating newly assembled 29,082 genomes from 845 fecal samples collected from three under-represented Asian countries—Korea, India, and Japan—the HRGM contains 232,098 non-redundant genomes of 5,414 representative prokaryotic species, >103 million unique proteins, and >274 million single-nucleotide variants. This is an over 10% increase from the largest reference database. The newly assembled genomes were enriched for members of the Bacteroidaceae family, including species associated with high-fiber and seaweed-rich diet. Single-nucleotide variant density was positively associated with the speciation rate of gut commensals. Ultra-deep sequencing facilitated the assembly of genomes of low-abundance taxa, and deep sequencing (>20 million read pairs) was needed for the profiling of low-abundance taxa. Importantly, the HRGM greatly improved the taxonomic and functional classification of sequencing reads from fecal samples. Finally, mapping homologous sequences for human auto-antigens onto the HRGM genomes revealed the association of commensal bacteria with high cross-reactivity potential with autoimmunity. The HRGM (www.mbiomenet.org/HRGM/) will facilitate the identification and functional analysis of disease-associated gut microbiota.

On a Non-Discrete Concept of Prokaryotic Species

The taxonomic concept of species has received continuous attention. A microbial species as a discrete box contains a limited number of highly similar microorganisms assigned to that taxon, following a polyphasic approach. In the 21st Century, with the advancements of sequencing technologies and genomics, the existence of a huge prokaryotic diversity has become well known. At present, the prokaryotic species might no longer have to be understood as discrete values (such as 1 or 2, by homology to Natural numbers); rather, it is expected that some microorganisms could be potentially distributed (according to their genome features and phenotypes) in between others (such as decimal numbers between 1 and 2; real numbers). We propose a continuous species concept for microorganisms, which adapts to the current knowledge on the huge diversity, variability and heterogeneity existing among bacteria and archaea. Likely, this concept could be extended to eukaryotic microorganisms. The continuous species concept considers a species to be delimited by the distance between a range of variable features following a Gaussian-type distribution around a reference organism (i.e., its type strain). Some potential pros and cons of a continuous concept are commented on, offering novel perspectives on our understanding of the highly diversified prokaryotic world, thus promoting discussion and further investigation in the field.

Evidence for selection in a prokaryote pangenome

A pangenome is the complete set of genes (core and accessory) present in a phylogenetic clade. We hypothesize that a pangenome’s accessory gene content is structured and maintained by selection. To test this hypothesis, we interrogated the genomes of 40 Pseudomonas genomes for statistically significant coincident (i.e. co-occurring/avoiding) gene patterns. We found that 86.7% of common accessory genes are involved in ≥1 coincident relationship. Further, genes that co-occur and/or avoid each other - but are not vertically or horizontally co-inherited - are more likely to share Gene Ontology categories, are more likely to be simultaneously transcribed, and are more likely to produce interacting proteins, than would be expected by chance. These results are not due to coincident genes being adjacent to one another on the chromosome. Together, these findings suggest that the accessory genome is structured into interacting sets of genes co-selected to function together within a given strain. Given the simi larity of the Pseudomonas pangenome with open pangenomes of other prokaryotic species, we speculate that these results are generalizable.

Sensing of autoinducer-2 by functionally distinct receptors in prokaryotes

Autoinducer-2 (AI-2) is a quorum sensing signal that mediates communication within and between many bacterial species. However, its known receptors (LuxP and LsrB families) are not found in all the bacteria capable of responding to this signaling molecule. Here, we identify a third type of AI-2 receptor, consisting of a dCACHE domain. AI-2 binds to the dCACHE domain of chemoreceptors PctA and TlpQ of Pseudomonas aeruginosa, thus inducing chemotaxis and biofilm formation. Boron-free AI-2 is the preferred ligand for PctA and TlpQ. AI-2 also binds to the dCACHE domains of histidine kinase KinD from Bacillus subtilis and diguanylate cyclase rpHK1S-Z16 from Rhodopseudomonas palustris, enhancing their enzymatic activities. dCACHE domains (especially those belonging to a subfamily that includes the AI-2 receptors identified in the present work) are present in a large number of bacterial and archaeal proteins. Our results support the idea that AI-2 serves as a widely used signaling molecule in the coordination of cell behavior among prokaryotic species.

Biological Functions of Prokaryotic Amyloids in Interspecies Interactions: Facts and Assumptions

Amyloids are fibrillar protein aggregates with an ordered spatial structure called “cross-β”. While some amyloids are associated with development of approximately 50 incurable diseases of humans and animals, the others perform various crucial physiological functions. The greatest diversity of amyloids functions is identified within prokaryotic species where they, being the components of the biofilm matrix, function as adhesins, regulate the activity of toxins and virulence factors, and compose extracellular protein layers. Amyloid state is widely used by different pathogenic bacterial species in their interactions with eukaryotic organisms. These amyloids, being functional for bacteria that produce them, are associated with various bacterial infections in humans and animals. Thus, the repertoire of the disease-associated amyloids includes not only dozens of pathological amyloids of mammalian origin but also numerous microbial amyloids. Although the ability of symbiotic microorganisms to produce amyloids has recently been demonstrated, functional roles of prokaryotic amyloids in host–symbiont interactions as well as in the interspecies interactions within the prokaryotic communities remain poorly studied. Here, we summarize the current findings in the field of prokaryotic amyloids, classify different interspecies interactions where these amyloids are involved, and hypothesize about their real occurrence in nature as well as their roles in pathogenesis and symbiosis.

Strategies for Optimizing the Production of Proteins and Peptides with Multiple Disulfide Bonds

Bacteria can produce recombinant proteins quickly and cost effectively. However, their physiological properties limit their use for the production of proteins in their native form, especially polypeptides that are subjected to major post-translational modifications. Proteins that rely on disulfide bridges for their stability are difficult to produce in Escherichia coli. The bacterium offers the least costly, simplest, and fastest method for protein production. However, it is difficult to produce proteins with a very large size. Saccharomyces cerevisiae and Pichia pastoris are the most commonly used yeast species for protein production. At a low expense, yeasts can offer high protein yields, generate proteins with a molecular weight greater than 50 kDa, extract signal sequences, and glycosylate proteins. Both eukaryotic and prokaryotic species maintain reducing conditions in the cytoplasm. Hence, the formation of disulfide bonds is inhibited. These bonds are formed in eukaryotic cells during the export cycle, under the oxidizing conditions of the endoplasmic reticulum. Bacteria do not have an advanced subcellular space, but in the oxidizing periplasm, they exhibit both export systems and enzymatic activities directed at the formation and quality of disulfide bonds. Here, we discuss current techniques used to target eukaryotic and prokaryotic species for the generation of correctly folded proteins with disulfide bonds.

All ANIs are not created equal: implications for prokaryotic species boundaries and integration of ANIs into polyphasic taxonomy

In prokaryotic taxonomy, a set of criteria is commonly used to delineate species. These criteria are generally based on cohesion at the phylogenetic, phenotypic and genomic levels. One such criterion shown to have promise in the genomic era is average nucleotide identity (ANI), which provides an average measure of similarity across homologous regions shared by a pair of genomes. However, despite the popularity and relative ease of using this metric, ANI has undergone numerous refinements, with variations in genome fragmentation, homologue detection parameters and search algorithms. To test the robustness of a 95–96 % species cut-off range across all the commonly used ANI approaches, seven different methods were used to calculate ANI values for intra- and interspecies datasets representing three classes in the Proteobacteria . As a reference point, these methods were all compared to the widely used blast-based ANI (i.e. ANIb as implemented in JSpecies), and regression analyses were performed to investigate the correlation of these methods to ANIb with more than 130000 individual data points. From these analyses, it was clear that ANI methods did not provide consistent results regarding the conspecificity of isolates. Most of the methods investigated did not correlate perfectly with ANIb, particularly between 90 and 100% identity, which includes the proposed species boundary. There was also a difference in the correlation of methods for the different taxon sets. Our study thus suggests that the specific approach employed needs to be considered when ANI is used to delineate prokaryotic species. We furthermore suggest that one would first need to determine an appropriate cut-off value for a specific taxon set, based on the intraspecific diversity of that group, before conclusions on conspecificity of isolates can be made, and that the resulting species hypotheses be confirmed with analyses based on evolutionary history as part of the polyphasic approach to taxonomy.