Hello everyone,
I am trying to predict miRNA target sites in long non-coding transcriptome. I have gone through the literature, but most of the algorithms such as TargetScan, miRanda, PicTar are focused on 3'-UTR of coding genes, which is not the case while working with the non coding genes.
Are there any other algorithms/tools which might help me in predicting the miRNA target sites in lincRNAs?
Or is there a way that I can use the above mentioned algorithms by modifying certain parameters?
You may try below tools to predict miRNA target sites in lncRNAs:
i) DIANA-LncBase: http://carolina.imis.athena-innovation.gr/diana_tools/web/index.php?r=lncbasev2%2Findex
DIANA-LncBase: experimentally verified and computationally predicted microRNA targets on long non-coding RNAs. Nucl. Acids Res. (2013) 41 (D1)
ii) miRcode: http://www.mircode.org
Respected Sir, thank you very much for your suggestions. I have already gone through the miRcode before.
1) The issues which I had with miRcode is that it uses an older version of GENCODE database as well as the human reference GRCh37. Furthermore, miRcode has not been updated after June 2012.
2)The Gene/transcript type of many of the lincRNAs have been changed to protein coding genes/transcripts in the current version. And since, miRcode has not been updated regularly, this might give rise to some false results.
In my case, I don't want to look for miRNA target sites in the databases. I have some human transcriptome samples with me and I'm interested in predicting miRNA target sites particularly in those samples.
Thank you very much for your suggestion. LncBASE v2 is a web based tool and it takes only one miRNA/lincRNA at a time.
I have a list of more than 100 differentially expressed miRNAs and lincRNAs with me, it would be really time consuming to search for each one of them individually. Furthermore, it would be little bit uncertain on my side to rely on the results of one algorithm.
However, I tried to do this work with MiRanda and RNAhybrid and I did find some good intersection between the results of EXPRESSED datasets. I even performed a Paerson correlation test to observe the functional enrichment of the expressed datasets and it did gave me interesting results.
So, it was very clear for me, whatever prediction algorithm we use, the best way to extract “good” interaction is to intersect with an expression dataset! and this might be because cells only express what they need for their regulations, and correlation analysis is a powerful tool to find biologically meaningful things.
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Nucleic Acids
- RNA
- Small RNA
- microRNA