I would love to hear from the community which is the best/most widely used tool in bio-informatics to distinguish between coding and non-coding RNA.
Lots of papers use CPAT, phyloCSF, and/or CPC2. They usually give decent overlap, so many people will use multiple of them. Based on the FEELnc paper, it seems like it performs just about as well as the others. They all have pretty good sensitivity, specificity, and accuracy.
(Update to the thread)
FEELnc is still a great choice, but some other great lncRNA annotation tools were published recently.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
In the FEELnc paper they are using some dataset and getting around 92% accuracy on a balanced 50/50 binary classification task. That seems to be about the usual performance for those algorithms, unless I am wrong. Would there be a great interest from the community for an algorithm that reaches say 94/95% or is it of little consequence given the current use?
Well, that really depends on your use case. Most papers that use these extensively are doing de novo lncRNA annotations, in which case a few extra % probably isn't going to mean much to them. However, micropeptides are beginning to emerge as products of transcripts traditionally thought of as non-coding, so I think tools to help distinguish those from conventional non-coding transcripts may be more useful than optimizing an algorithm for 1-2% greater accuracy.
Thanks for the insight. Coming from a CS background I am working on a neural network approach to lncRNA detection. Ideally would like to find co-authors for a paper in this area. Probably a topic for another post.