Accuracy in Predicting Secondary Structure of Ionic Channels

Ionic channels are among the most difficult proteins for experimental structure determining, very few of them has been resolved. Bioinformatical tools has not been tested for this specific protein group. In the paper, prediction quality of ionic channel secondary structure is evaluated. The tests were carried out with general protein predictors and predictors only for transmembrane segments. The predictor performance was measured by the accuracy per residue Q and per segment SOV. The evaluation comparing ionic channels and other transmembrane proteins shows that ionic channels are only slightly more difficult objects for modeling than transmembrane proteins; the modeling quality is comparable with a general set of all proteins. Prediction quality showed dependence on the ratio of secondary structures in the ionic channel. Surprisingly, general purpose PSIPRED predictor outperformed other general but also dedicated transmembrane predictors under evaluation.

• Type: Conference
• Year: 2009
• Where: ICCCI
• Authors: Bogumil Konopka, Witold Dyrka, Jean-Christophe Nebel, Malgorzata Kotulska
• Keywords: ICCCI 2009, Internet Technology, Ionic Channel, Proteins, Transmembrane