Biomedical Engineering Reference
In-Depth Information
4.6 Discrimination of Different Classes of Ion Channel Proteins
Ion channels are integral membrane proteins that enable the passage of inorganic
ions across cell membranes. They are key components for physiological functions
and the evidence of ion channels for the role in diseases has been described in a
special issue in Progress in Biophysics and Molecular Biology [
85
]. It has been
mentioned that different types of ion channels voltage-gated potassium, calcium,
sodium, and ligand-gated channels perform different functions. Hence, several
methods have been proposed to discriminate the ion channels and classifying
them into different groups. Saha et al. [
86
] developed a method based on support
vector machines to discriminate ion channels and nonion channels and reported an
accuracy of 89% to discriminate them. Further, the ion channels have been classi-
fied into potassium, sodium, calcium, and chloride channels with one against others
and obtained an average accuracy of 97.8%. A web server, VGIchan has been
developed for predicting and classifying voltage-gated ion channels, and it is
available at
www.imtech.res.in/raghava/vgichan/
.
Recently, Lin and Ding [
87
] utilized a feature selection technique, analysis of
variance and support vector machines to detect ion channels and classifying them.
They showed an accuracy of 86.6% for discriminating ion channels and nonion
channels. Further, voltage- and ligand-gated channels are distinguished with an
accuracy of 92.6% and four types of channels (potassium, sodium, calcium, and
anion) are classified with an accuracy of 87.8%.
5 Drug-Target Interactions
The identification of molecular target is a critical step in drug discovery and
development. Ion channels are one of the most popular drug targets in various
diseases including cardiovascular and central nervous systems. Hence, several
bioinformatics methods have been developed to predict the potential drug targets
with ion channels and other proteins. Yamanishi et al. [
88
] characterized four
classes of drug-target interaction networks in humans involving enzymes, ion
channels, G-protein coupled receptors and nuclear receptors and revealed signifi-
cant correlations between drug structure similarity, target sequence similarity and
the drug-target interaction network topology. The information on interaction
between drugs and target proteins has been obtained from KEGG [
89
], BRENDA
[
90
], SuperTarget [
91
], and DrugBank databases [
92
]. Further, they developed
statistical methods to predict unknown drug-target interaction networks from
chemical structure and genomic sequence information simultaneously on a large
scale.
He et al. [
93
] developed the datasets for drug-target pairs (positive dataset) and
non drug-target pairs (negative dataset) and utilized them for predicting
drug-target interactions. The positive dataset contains a total of 4,797 drug-target
