Using Hilbert-Huang transform, subcellular localization tbr prokaryotic and eukaryotic proteins was predicted and tested with Reinhart and Hubbard＇s dataset. The prediction accu- racy of prokaryotic and eukaryotic pr...Using Hilbert-Huang transform, subcellular localization tbr prokaryotic and eukaryotic proteins was predicted and tested with Reinhart and Hubbard＇s dataset. The prediction accu- racy of prokaryotic and eukaryotic protein sequences concentrated in the dataset all reached 100% by self-consistency, 91.8% for the former and 88% for the latter by the five fold cross-validation test. A significant improvement in prediction quality by incorporating the spectrum parameters with the conventional amino acid composition was observed. One of the crucial merits of this approach is that many existing tools in mathematics and engineering can be easily applied in the predicting process. It is anticipated that digital signal processing may serve as a useful vehicle for many other protein science areas.展开更多
基金Supported by the Fundamental Research Founds for the Central Universities(2010JC003)
文摘Using Hilbert-Huang transform, subcellular localization tbr prokaryotic and eukaryotic proteins was predicted and tested with Reinhart and Hubbard＇s dataset. The prediction accu- racy of prokaryotic and eukaryotic protein sequences concentrated in the dataset all reached 100% by self-consistency, 91.8% for the former and 88% for the latter by the five fold cross-validation test. A significant improvement in prediction quality by incorporating the spectrum parameters with the conventional amino acid composition was observed. One of the crucial merits of this approach is that many existing tools in mathematics and engineering can be easily applied in the predicting process. It is anticipated that digital signal processing may serve as a useful vehicle for many other protein science areas.
