Agnodice: indexing experimentally supported bacterial sRNA-RNA interactions

Abstract

In the last decade, the immense growth in the field of bacterial small RNAs (sRNAs), along with the biotechnological breakthroughs in Deep Sequencing permitted the deeper understanding of sRNA-RNA interactions. However, microbiology is currently lacking a thoroughly curated collection of this rapidly expanding universe. We present Agnodice (https://dianalab.e-ce.uth.gr/agnodice), our effort to systematically catalog and annotate experimentally supported bacterial sRNA-RNA interactions. Agnodice, for the first time, incorporates thousands of bacterial sRNA-RNA interactions derived from a diverse set of experimental methodologies including state-of-the-art Deep Sequencing interactome identification techniques. It comprises 39,600 entries which are annotated at strain-level resolution and pertain to 399 sRNAs and 12,137 target RNAs identified in 71 bacterial strains. The database content is exclusively experimentally supported, incorporating interactions derived via low yield as well as state-of-the-art high-throughput methods. The entire content of the database is freely accessible and can be directly downloaded for further analysis. Agnodice will serve as a valuable source, enabling microbiologists to form novel hypotheses, design/identify novel sRNA-based drug targets, and explore the therapeutic potential of microbiomes from the perspective of small regulatory RNAs.IMPORTANCEAgnodice (https://dianalab.e-ce.uth.gr/agnodice) is an effort to systematically catalog and annotate experimentally supported bacterial small RNA (sRNA)-RNA interactions. Agnodice, for the first time, incorporates thousands of bacterial sRNA-RNA interactions derived from a diverse set of experimental methodologies including state-of-the-art Next Generation Sequencing interactome identification techniques.

Keywords: RNA regulation; bacteria; interactions; microbiology; small RNAs.

Fig 1

Schematic representation of the development and information flow, from the researcher bench to the creation of Agnodice resource. sRNA interactome data, generated by low- or high-yield experimental methods, are contained in hundreds of articles and Supplementary Materials. After querying for candidate sources, meticulous curation is applied. sRNA-RNA interactions are detected by means of manual curation. The resulting set of entries is harmonized, accompanied with a rich experiment- or study-specific metainformation and broken down into an efficient database schema. Inter-connection with external resources, including PubMed and NCBI Taxonomy and Peryton database of microbe-disease associations, is performed. The resulting content is provided in the form of an open, user-friendly online database, supporting numerous querying, filtering, visualization, and download functionalities.

Fig 2

(a) Top 12 most frequent sRNA regulators and (b) regulated genes featured in Agnodice. (c) Top eight species in terms of total interactions in the database. (d) Number of total interactions per experimental method (interactions by low-yield methods summed together). Interaction sums are transformed in log10 space. As expected, these numbers are potentially skewed toward sRNAs originating from organisms that are extensively studied (e.g., E. coli).

Fig 3

Demonstration of the results offered through the user interface of Agnodice as generated by an example query. Basic information is given in a first-level tabular format, while supplemental information regarding each initial entry can be disclosed through specialized buttons that allow users to reveal the corresponding sub-tables hosting it.