Available Software
miRNATIP: A self-organizing map based miRNA-Target Interactions Predictor
Motivation
MicroRNAs (miRNAs) are small non coding RNAs with regulatory functions to post-transcriptional level. They play an important role in molecular and cellular mechanisms, thanks to their ability to bind and interact with many RNA messengers (mRNAs) of coding product involved in a wide range of biological pathways, cellular status, and conditions. They are more often seen as key regulators of pathological conditions as cancer. Indeed a wide range of studies evidences a potential role of these small molecules as biomarkers in this disease. The first evidence of miRNAs arose in Caenorhabditis elegans (Nematoda; Rhabditidae). Studies on nematodes have been applied in a wide range of fields, in order to understand the molecular biology of humans and animals. In animals, the miRNA-mRNA interaction occurs through an imperfect binding, which makes the computational prediction of the target a challenging task. We present here a Self Organizing Map (SOM) based method for the miRNA target prediction, characterized by the projection of any mRNA sequence allowing to find an interaction with miRNAs.
Methods
The miRNA target prediction method we propose is composed of four steps and it is shown in the figure. In the first step, a set of miRNAs is used for the training of a SOM. In detail, in this step, we consider only the 8-mer miRNA seeds, because it has been demonstrated the seed is mainly responsible for the miRNA target binding. Each miRNA seed is then converted in a 4X8 (nucleotide symbols vs. seed length) Position Weight Matrix (PWM), according to the representation used for motifs in biological sequences. At this point, a SOM is trained with this set of PWMs. The second step is the projection of a mRNA sequence over the trained SOM. For this reason, we extracted all the 8-length mRNA fragments through a 8-mer sliding window. This way, we obtained a set of 4x8 PWMs that can be projected over the trained SOM. The result of this step is, for each neural unit (cluster), a list of miRNA_seed-mRNA_fragment. Each cluster can be considered as a preliminary list of predicted miRNAs-mRNAs interactions. In the third step, we filtered these putative interactions considering the remaining part of the miRNA sequences (miRNA tail). For each miRNA_seed-mRNA_fragment interaction, we considered respectively the miRNA tail and the mRNA sequence of the same length of miRNA tail, next to the predicted mRNA_fragment. Then we computed a dissimilarity measure based on normalised euclidean distance between the PWM representations of those two sequences, and we retained only the couples of miRNA-mRNA interactions whose distance is below a certain threshold. In order to take into account also the presence of possible bulge loops between the 8-mer seed and the tail of the miRNA, we considered an offset for the selection of the mRNA fragment corresponding to the miRNA tail. Finally, in the fourth step, we performed another filtering to the miRNA-mRNA interaction list, by computing the free-energy of the miRNA-target site duplex.
Results
We tested our method by predicting the miRNA target interactions of the C. elegans species. We considered all the 434 miRNA mature sequences found in miRBase, release 21, and the 3' UTR mRNA sequences found in WormBase , release WBcel 235. As gold standard, we used the 3209 validated interactions provided by miRTarBase repository, release 4.5. We compared our results with other two target predictors: PITA and miRanda. After different runs, we set network size to 30x30 neurons, free-energy threshold = -17 kcal/mol, miRNA_tail-mRNA_fragment distance threshold = 0.7, offset = 3. For the test procedure, we considered only the 3' UTR mRNA having at least a validated target. With this configuration, we obtained 106803 target predictions, PITA and miRanda provided 135880 and 64333 predictions respectively. With regards to the validated targets, we reached an accuracy of 70.6%, whereas PITA and miRanda obtained an accuracy of 61.1% and 26.9%.
Figure 1 - miRNATIP pipeline.
Here the miRNATIP web-application.
How to Cite
MiRNATIP: a SOM-based miRNA-target interactions predictor,
Fiannaca A., La Rosa M., La Paglia L., Rizzo R. and Urso A.,
BMC Bioinformatics (Suppl 11):321, 2016,
doi: https://doi.org/10.1186/s12859-016-1171-x
Source Code
miRNATIP has been developed with Java Platform, Standard Edition 8. A script in Phyton version 2.7 is used in order to easily execute the pipeline in Fig.1. miRNATIP also requires the IntaRNA tool v. 1.2.5.
miRNATIP source code can be download here. IntaRNA tool can be download here.