MiRNA Text Mining in Abstracts.

miRetrieve

miRetrieve is designed for microRNA text mining in abstracts. By extracting, counting, and analyzing miRNA names from literature, miRetrieve aims at providing biological insights from a large amount of text within a short period of time.

Getting Started

An online version with the most important functions of miRetrieve is available under https://miretrieve.shinyapps.io/miRetrieve/.

To install miRetrieve from CRAN, run

install.packages("miRetrieve")

Alternatively, you can also install miRetrieve from GitHub by running

install.packages("devtools")

devtools::install_github("JulFriedrich/miRetrieve",
        dependencies = TRUE,
        repos = "https://cran.r-project.org/")

miRetrieve is built around the idea of using field-specific PubMed abstracts from PubMed to characterize and analyze microRNAs in disease-related fields (e.g. "miRNAs in diabetes").

To get started, download a microRNA-related abstract from PubMed via Save - Format: PMID - Create file and load it into R using

df <- miRetrieve::read_pubmed("PubMed_file.txt")

and subsequently extract all microRNAs with

df <- extract_mir_df(df)

An extensive Vignette with the underlying mechanism, functions, and a complete workflow is available under

https://julfriedrich.github.io/miRetrieve/articles/miRetrieve.html

Authors

Julian Friedrich, Hans-Peter Hammes, Guido Krenning

License

miRetrieve is published under the GPL-3 license.

Publication

miRetrieve and its functions are presented in a manuscript, currently under review.

Supplementary Files referenced in the manuscript are located in a different repository, freely available under

https://github.com/JulFriedrich/miRetrieve-paper

Reference

Acknowledgments

• join_mirtarbase is based on the latest miRTarBase version 8.0 (http://miRTarBase.cuhk.edu.cn/). If you use miRetrieve to visualize miRNA-mRNA interactions based on miRTarBase, please make sure to cite Hsi-Yuan Huang, Yang-Chi-Dung Lin, Jing Li, et al., miRTarBase 2020: updates to the experimentally validated microRNA–target interaction database, Nucleic Acids Research, Volume 48, Issue D1, 08 January 2020, Pages D148–D154, https://doi.org/10.1093/nar/gkz896.

• compare_mir_terms_log2(), compare_mir_count_log2(), and compare_mir_terms_scatter() are greatly inspired by “tidytext: Text Mining and Analysis Using Tidy Data Principles in R.” by Silge and Robinson (https://www.tidytextmining.com/). In addition, "tidytext" provides a valuable resource of general text mining in R.

• Key packages for miRetrieve are tidytext, topicmodels, and the packages included in the tidyverse (see Vignette).