Detection of miRNAs is a difficult problem. Due their small size limits the
available information and current sensitive methods, such as: blast, nhmmer,
or cmsearch are designed to increase sensitivity, but lead to an inevitable
large number of false positives only detected by detailed analysis of specific
features of typical miRNAs and/or conservation patterns in a structure-annotated
multiple sequence alignments.
The miRNAture pipeline implements a workflow specific to animal microRNAs
that automatizes homology search and validation steps.
On the homology search it combines two modes: sequence-homology by blast or/and
nhmmer using query sequences or hidden markov models (HMMs), and structural
validation performed by the INFERNAL package, using covariance models (CMs).
A merging step produces a final list of homology candidates. Over those
candidates a Mature annotation stage performs a correction of the position of
mature sequences on the detected precursor and a structural evaluation
in terms of minimum free energy (MFE), precursor length, folding and the
evaluation of anchored family specific-multiple secondary alignment
(using MIRfix). Final sanity checks are performed on the Evaluation stage,
that reviews all the last mature annotation process, filtering the invalid candidates
at structure level and reporting valid candidates on GFF3/BED and fasta files
together with a summarize file that provides overall information about detected
miRNA candidates and families.
The easiest way to install miRNAture is through conda. To do so, please first install
conda.
To speed up installation of dependencies and packages we suggest to use mamba, for this just run:
conda install mamba -c conda-forge
You can use mamba as drop-in replacement for conda by simply replacing the
call to conda with a call to mamba.
To install miRNAture from conda in a specific mirnature environment
simply run:
mamba create -n mirnature mirnature
if mamba is available, else run:
conda create -n mirnature mirnature
Create a mirnature conda environment with the file miRNAture.yml:
mamba env create -n mirnature -f miRNAture.yml
Activate the environment containing all dependencies:
conda activate mirnature
followed by the manual steps:
perl Build.PL
./Build
./Build test
./Build install
which will install miRNAture in the mirnature conda environment.
The most important input file is a DNA sequence. This could be a multi-fasta
sequence that belongs from a common specie (i.e. complete genome or group of
particular sequences). At the same time, previous to execute miRNAture a
pre-calculated dataset (that contains default data as CMs, HMMs, and required
files to perform mature prediction) must be downloaded and correctly indicated
in the command line options with the flag -dataF.
New in version 1.1 A new dataset containing all miRBase HMMs/CMs and validated mature sequences is recommended to use as first approach to identify miRNAs over target species. This dataset can be downloaded from here.
To run miRNAture in its complete mode with default options, just run as:
# Activate the mirnature environment
conda activate mirnature
# Run miRNAture
./miRNAture -stage complete -dataF <Precalculated_folder> -speG <Target Genome> -speN <Specie_name> -speT <Tag_specie> -w <Output_dir> -m <Mode> (-str <Blast_strategy>) -blastq <Blast_queries_folder>
Final predicted miRNAs will be written on the <Output_dir> indicated with the -w flag.
The final candidates are described on the folder Final_miRNA_evaluation/ as
follows:
Final_miRNA_evaluation/
├── Fasta/
├── MFE/
├── miRNA_annotation_Lach_accepted_conf.bed
├── miRNA_annotation_Lach_accepted_conf.gff3
├── miRNAture_summary_Lach.txt
└── Tables/
Inside this folder, miRNAture will create 3 folders containing their
correspondent results: sequences in fasta format (Fasta/), minimum free
energy and lengths from described sequences (MFE/) and the supporting
information ordered in tables for each annotated candidate (Tables/).
Additionally, associated genomic positions for the miRNA candidates are reported
in BED and GFF3 formats and a summary file, miRNAture_summary_*.txt, that
describes overall descriptive statistics from found miRNA families.
For detailed instructions how to use miRNAture please refer to the Manual pages:
- Through your favourite explorer, open the manual pages here.
Pre-calculated data composed by miRNA CMs, HMMs and required input files to perform mature annotation has to be downloaded before run the full miRNAture pipeline. Available datasets are listed below: