|
Specifically:
In the prediction model PSKAcePred, the sequences fragments are firstly
extracted in window size -10 to +10 (acetylation lysine centered in position of
0). Then, 13 optimization positions are chosen by using position-specific method.
Users can submit protein sequence(s) in FASTA format to this web interface.
The system efficiently returns the prediction results, including protein name,
the position of site, flanking amino acid sequences and SVM probability. In the
output of flanking amino acid sequences, the amino acids in green are those selected by
using information gain and the red amino acids (K) are predicted as acetylation
sites.
There is no upper bound for the protein sequence length, input protein sequences
have to: (i) contain minimum 21 amino acids; (ii) contain only characters that
present amino acids.
Maximum 20 protein sequences is recommend to be submitted in the textbox. Too
many sequences for prediction can cause the system to crash. To carry out
large-scale predictions, the researchers can download the Matlab codes about the
PSKAcePred below.
Download:
The training data can be downloaded at
here. The independent data can be downloaded at
here. The Matlab codes can be downloaded at
here .
|