The package hosts variable databases containing miRNA-target interactions, and provides functions to communicate with the multiMiR web server and its database.
BiocManager::install("multiMiR")
- We need a list of miRNAs' name and a list of targets, such as genes and lncRNAs, to identify the interactions. miRNA-gene and miRNA-lncRNA interactions will be defined.
miRNAs are known as source nodes, while genes and lncRNAs are known as targets
The function get_multimir
provides the interactions we need based on
hsa; homo sapiens project.
-
we can choose the
predicted.cutoff
argument in this function to define the percentage of interactions. Usually, 80 prediction cutoff means top 80% significant interactions based on validated databases including miRTarBase, miRDB, TargetScan, and TarBase. We can use 20 cutoff in a highly significant detection of results, exploring less interactions number than 80 cutoff. -
Argument
table
in the function means the generation of table header and footer. It usually is "all". -
Argument
predicted.cutoff.type
means the type of predicted cutoff is used. Usually "P" as percent is used.
<- get_multimir(org = "hsa",mirna = miRNA, target = genes, predicted.cutoff = 80, table = "all",predicted.cutoff.type = "p")
- You can filter the final output table based on database column to release the results based on the validated databases.
database mature_mirna_acc mature_mirna_id target_symbol target_entrez target_ensembl experiment support_type pubmed_id type score
mirtarbase MIMAT0004954 hsa-miR-543 ADD1 118 ENSG00000087274 HITS-CLIP Functional MTI (Weak) 23824327 validated NA
mirtarbase MIMAT0002818 hsa-miR-496 AKT1 207 ENSG00000142208 PAR-CLIP Functional MTI (Weak) 22100165 validated NA
mirdb MIMAT0004954 hsa-miR-543 FBXO34 55030 ENSG00000178974 NA NA NA predicted 99.51948263
mirdb MIMAT0004954 hsa-miR-543 NWD2 57495 ENSG00000174145 NA NA NA predicted 99.49892675
targetscan MIMAT0004954 hsa-miR-543 CASK 8573 ENSG00000147044 NA NA NA predicted -0.156
targetscan MIMAT0004954 hsa-miR-543 CMPK1 51727 ENSG00000162368 NA NA NA predicted -0.156
tarbase MIMAT0004954 hsa-miR-543 DHX33 56919 ENSG00000005100 Degradome sequencing positive validated NA
tarbase MIMAT0004954 hsa-miR-543 ADAM22 53616 ENSG00000008277 Degradome sequencing positive validated NA
-
Score column shows the significant interactions in two databases including mirdb and targetscan, which means that we can filter the interactions with 80% cutoff only, describing more than 80 percent significant interactions. The negative scores defined by targetscan database based on the negative correlation of interactions and inhibitory effects of miRNA on target, which means highly negative scores, more significant effects of miRNA on targets. So we can choose -0.5 or -0.8 cutoff as a score criteria to filter targetscan results.
-
type column shows the type of interactions and database function. predicted or validated results. We can choose one or both of them.
-
For the databases without score number to choose, we can filter the support-type column, so that we remove weak functional MIT and keep positive, strong, or those written only functional MIT. MIT: MicroRNA-Target Interaction
You should add your attribute features to your final interactions table. Any feature describing the targets attributes. Clinical features, Molecular identification, cellular attribute, regulation or mutation levels, treated or non-treated states, or etc can be used. These features are used for detecting and defining particular targets in the ceRNA network.
- The final prepared table is transferred to Cytoscape software for contructing and analyzing the ceRNA network. Source nodes should be defined as miRNA column with miR-Ids, and Target nodes should be defined as target columns with symbols or ids. The columns related to attributes should be defined as target-attribute. Then, You can design the network in Cytoscape.