mirPRo / Blog: Recent posts

This feature is enabled after versoin 1.1.2 (needs gcc and g++ version 4.9+).

If your BAM files come from read-to-genome alignment, you need to convert the BAM files to FASTA files first, and then map the FASTA files to percursor miRNA sequences in miRBase ("hairpin.fa"). The new BAM/SAM alignment files can be processed in the following steps.

If your BAM files come from read-to-precursor-miRNA alignment, they can be processed directly in the follwing steps.... read more

A.

• Web links of datasets (used in our paper) to test our program:

Reference sequences:
mouse: ftp://ftp.ensembl.org/pub/release-81/fasta/mus_musculus/dna/Mus_musculus.GRCm38.dna.toplevel.fa.gz
human: ftp://ftp.ensembl.org/pub/release-81/fasta/homo_sapiens/dna/Homo_sapiens.GRCh38.dna.primary_assembly.fa.gz

Gene annotation files:
mouse: ftp://ftp.ensembl.org/pub/release-81/gtf/mus_musculus/Mus_musculus.GRCm38.81.gtf.gz
human: ftp://ftp.ensembl.org/pub/release-81/gtf/homo_sapiens/Homo_sapiens.GRCh38.81.gtf.gz... read more

A: You can go to directory "result/XXX/", where "XXX" is the sample name, and find the files "XXX_total_mapping.csv" and "XXX_mature_miRNA_mapping.csv". The first file shows the read mappings for all reads that can be mapped to the known precursor miRNAs; while the second file shows the read mappings for the reads that are considered as mature miRNA reads. How to determine mature miRNA reads? Please read our publication ( http://www.nature.com/articles/srep14617 ) for details. Both files inlude the read alignments for all known miRNA precursors.

A: You can use many tools for this purpose. For instance, we have used DESeq2. Since these statistical tools have normalization prcedure, it is important for users to combine both novel and known miRNA data for the downstream differential expression analysis. Our recomendation is following:

(1) For differential analysis of mature miRNA, you should combine both resultant files (i.e., result_mature.csv and result_novel_mature.csv). ... read more

A: We first map known mature miRNAs to knwon pre-miRNAs to get the annotation (5' arm and 3' arm) for each known pre-miRNA; then map the clean collapsed reads to known pre-miRNAs, filter them according to our alignement criteria, and assign the read count to the corresponding pre-miRNA arm (if reads have non-unique alignements, select the alignment(s) with minimum soft clip numbers and then divide the count evenly). Finally, for each pre-miRNA, we get the read counts for 5' and 3' arms respectively, and we choose the arm with higher read count as the mature arm, and consider the read count of mature arm as the expression value of this pre-miRNA (or miRNA gene). If both read counts for both arms are the same, we randomly pick one as the expression value for the given pre-miRNA. The miRNA precusor expression data can be found in the file In the result file: "result/result_precursor.csv". We recommend users to use this result for down stream analysis.

A: In the result file: "result/result_mature.csv", we provide the read counts for each mature miRNA annotated in miRBase of a given species in your samples. These counts can be considered as the expression values for mature miRNAs. In some situations, one mature miRNA can be generated from different precursors locating in different loci (or different chromsomes), so the expression value of this mature miRNA can't represent the expressin value of a single precursor miRNA. In other words, we can't simply use the mature miRNA counts to represent the miRNA gene expression level. What is the solution? See Q.3.

A: For a given read, it can be often mapped to the forward strand of some precoursors. Because some precursor miRNAs can be partialy reverse complementary to each other. Therefore, this read can also be mapped to the reverse strand of some precursors. Like other miRNA analysis tools (e.g. miRDeep2), we only focus on the mappings that are mapped to the forward strands of precuors and filter out those mappings mapped to the reverse strands of precursors.