TL;DR
This paper introduces RPRH, a novel web-based tool that predicts the potential roles of plant microRNAs (miRNAs) in the human body. By integrating diverse datasets and employing a robust computational framework, RPRH aims to bridge the gap in understanding cross-kingdom miRNA regulation, a burgeoning field with significant implications for medicine and drug discovery.
What Problem Does This Paper Address?
Recent research has revealed that plant-derived miRNAs can be absorbed by animals and influence their physiological processes. This "cross-kingdom regulation" opens exciting avenues for miRNA research and drug development. However, a significant bottleneck has been the lack of specialized bioinformatics tools to systematically investigate these interactions. Existing tools are not designed to handle the unique challenges of predicting plant miRNA functions in a human context, leaving a critical gap in our understanding of how these foreign molecules might impact human health.
Methods
The authors developed a multi-step computational pipeline to infer potential plant miRNA roles in humans:
- Data Collection: They gathered mature plant miRNA sequences from miRBase and human 3'UTR sequences from UCSC and TargetScan. Functional gene sets, including disease, pathway, and gene ontology (GO) annotations, were sourced from GeneSetDB, MsigDB, and literature.
- Plant miRNA Target Prediction: To predict which human genes are targeted by plant miRNAs, they employed three established miRNA target prediction tools: TargetScan, miRanda, and PITA. The consensus prediction from these tools was used to define the target gene set for each plant miRNA.
- Gene Set Enrichment Analysis: Using a hypergeometric test, they performed enrichment analysis on the predicted target genes against the curated functional gene sets. This step aims to identify biological pathways, diseases, or functions that are significantly associated with the targets of a given plant miRNA.
- Endogenous miRNA Similarity Calculation: To assess the novelty of plant miRNA functions, they calculated the Jaccard similarity between the target gene sets of plant miRNAs and human endogenous miRNAs. This helps identify plant miRNAs that might exert unique effects in humans due to a lack of similar endogenous counterparts.
Datasets
The study utilized the following datasets:
- Plant miRNA sequences: Downloaded from miRBase (release 22).
- Human 3'UTR sequences: Obtained from the UCSC genome browser (GRCh38) and the TargetScan website.
- Functional gene sets: Sourced from GeneSetDB, MsigDB, and relevant literature.
- Experimentally validated human miRNA targets: Downloaded from miRTarBase.
The paper does not introduce new benchmark datasets but rather integrates existing resources to build its comprehensive database.
Key Insights and Ideas
- Cross-kingdom regulation is a vast and underexplored landscape: The sheer number of plant miRNAs and their potential targets highlights the significant, yet largely uncharacterized, impact plant miRNAs could have on human biology.
- RPRH provides a crucial bridge: By developing a dedicated tool, the authors enable researchers to systematically explore this complex interaction space, moving beyond anecdotal evidence.
- Functional annotation is key to understanding roles: Simply predicting targets isn't enough; linking these targets to known biological functions (pathways, diseases, etc.) is essential for inferring the roles of plant miRNAs.
- Novelty matters: Identifying plant miRNAs that lack similar human counterparts is important, as these may represent unique therapeutic targets or sources of novel biological insights.
Key Results
- Comprehensive Database: RPRH integrates data from 10,414 mature plant miRNAs across 82 plant species and 18,062 annotation terms (diseases, pathways, GO, drugs).
- Extensive Annotations: The analysis yielded 40,743,425 miRNA-functional annotation associations, with most miRNAs having less than 5000 annotations.
- Known Regulation Captured: The method successfully identified known cross-kingdom regulatory roles, such as osa-miR168a-5p's involvement in glucose metabolism and gma-miR159a-3p's enrichment in WNT signaling pathways.
- Identification of Novel Players: A significant number of plant miRNAs (2390) were found to have no similar endogenous human miRNAs, suggesting they might perform distinct roles in the human body.
- Web Server Availability: The RPRH web server is publicly available at http://www.rnanut.net/rprh/, providing a user-friendly interface for searching and exploring these predictions.
Limitations and Future Work
- Reliance on Mammalian Prediction Tools: The study acknowledges that it uses tools developed for mammalian miRNA target prediction. It notes that if special rules or mechanisms for cross-species miRNA-target regulation are discovered in the future, novel algorithms might be needed.
- Unclear Mechanisms: The authors state that it remains largely unclear how plant miRNAs regulate human genes, highlighting this as an area for future research.
- Experimental Validation Needed: While the tool provides predictions, experimental validation is crucial to confirm the inferred roles of plant miRNAs in human bodies.
- Algorithm Refinement: Future work could focus on developing more specialized algorithms that account for the unique characteristics of plant-miRNA-human interactions.
Summary updated