| 摘要 |
The present invention relates to a method for predicting a protein-binding base in an RNA sequence and, more particularly, to a novel method for predicting protein-binding sites in the RNA sequence with consideration for RNA binding partners. A support vector machine (SVM) model, which predicts an RNA base binding with protein by using both of RNA and protein sequences, is developed and the SVM model, which predicts a base binding with protein by using only the RNA sequence, is developed for comparison. As a result, a prediction by using both of the RNA and protein sequences is confirmed to be superior to a prediction by using the RNA sequence only in most evaluation tests. According to the new model, when the binding partners are changed, binding sites, which are different from each other with respect to a given RNA sequence, are predicted. Therefore, the base of an RNA binding with protein is predictable by using only sequence data even if a structure thereof is unknown, thus the prediction is usefully applicable to biochemical experiments. |
