Dear all
I need a help understanding this please I am new to long non coding RNA and my background is computational.
When I run a lncRNA tool to detect a novel lncRNA, it finds a lncRNA x for example. when I look this x up in public data bases it is annotated as protein coding.
Does this mean the tool has a new evidence that makes x a lncRNA not protein coding?
And most tools runs per trancsript. When I annotate the transcripts, we can have a gene found say in 5 transcripts. 2 of those transcripts are predicted as non coding and the rest as coding. can I safely say, this gene is a lncRNA, or all the transcripts that the gene span have to be non coding to say the gene is a lncRNA? The goal is to find differentially expressed lncRNA at some point.
Thanks
You need to be careful and look at the underlying data. If there is experimental evidence to support that then it does not matter what a computer program says ... the experimental evidence wins. There is a slim chance the experimental data is wrong but if it is human curated (and has multiple independent instances of evidence) then it should be trustworthy.
Good point too Genomax, I can check this too ! Thanks so much :)
Hey again Sharon, how did you identify the lncRNA? Just because some program identifies it as such is not proof, in my opinion, that it may never be translated into a protein or protein domain. It may lack the key sequences that allow for translation to happen, but you have not mentioned which program you've used, so, I don't know if the program takes such information into account.
Hi Kevin Thanks so much. I used lncScore https://www.nature.com/articles/srep34838.pdf?origin=ppub how do you think?
Hi Sharon, I have to say that I am a fan of Kai Wang and he has a good reputation in the field. I was not aware of this manuscript by his group but it seems pretty convincing. For starters, the tool uses HISAT2 / StringTie (or Trinity de novo assembler) and then appears to predict a transcript's potential of being a lncRNA based on a model that they have produced from current known lncRNAs, plus other key features.
You likely have a score assigned to each of your transcripts, indicating the likelihood that it is a lncRNA.
Great thanks, as a start I ran that using tophat cufflink transcripts we already have (just to have a quick sense of water). I am having another pipeline using Hisat is running to get better transcripts to use with lncScore. Thanks a lot for your feedback Kevin. Much appreciated as usual.