While doing protein intrinsic disorder prediction, is it requisite to take only reviewed sequences from the proteome or we can proceed with unreviewed also?Since i have gone through the papers based on disorder prediction and it seems that mostly reviewed sequences has been taken for the analysis.
The un reviewed sequences are generated from "just" computer software without any manual (human) level validation because there may not be hard data such as N-terminal sequencing to support the prediction. This is good as you are working with a un-explored protein sequence. If you base your disorder prediction upon another prediction the results may not be reliable. The best solutions is to start with verifying the sequences by mass spec like techniques or support your end prediction with hard data! I hope this helps...
You can enter any sequence into the appropriate disorder prediction servers, e.g. http://dis.embl.de, and they will return a prediction of disorder probability for the sequence you entered. Of course, to get a meaningful answer, your sequence has to appropriately represent the query protein, which for a genomic sequence also involves the accurate prediction of splicing sites and reading frame. A reviewed sequence has been verified by collecting the data from multiple sources, looking for discrepancies, and pre-processed with a variety of prediction tools. If your enter your own sequence, it is up to you to make sure it accurately represents the protein you wish to analyze. As with any computational method, the GIGO law applies: garbage in, garbage out.
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology