NETASA:
Neural Network Application to Predict Solvent Accessibility of Proteins
This program predicts solvent accessibility of residues in a given
amino acid sequence. Predictions are made according to our new algorithm, based on
neural network, which
has been trained for different accessibility states using various
threshold values. Certain choices are presented here. A choice
of threshold means that a residue was classified into exposed (e),
buried (b) or partially buried (p) states, depending on the value of
its accessible surface area falling within the range described therein.
e.g. If a user chooses 25% threshold then a residue predicted to be
buried means it has less than 25% surface area accessible to solvents.
Thus in a 25% choice, residues are classified as "buried" or "exposed", if they
have less or more than 25% exposed surface area, respectively. If a user chooses
25-75% case, it means that the states are classified as:
Buried: 0-25% exposed surface area.
Partially Buried: 25-75% exposed surface area.
Exposed: More than 75% exposed surface area.
While submitting your queries, please note the following:
1. Enter sequence in one letter code without spaces.
2. Carriage return for next line are allowed but spaces are not.
3. Largest sequence size is 700. If you indeed have a larger sequence to predict, contact
Shandar Ahmad
***************************************************************************
Reference: Shandar Ahmad and M. Michael Gromiha, Bioinformatics (2002) 18, 819-824 .
***************************************************************************