3 <title>JNet Secondary Structure Prediction</title>
\r
6 <strong>JNet Secondary Structure Prediction</strong>
\r
8 Secondary structure prediction methods attempts to infer the likely secondary
\r
9 structure for a protein based on its amino acid composition and
\r
10 similarity to sequences with known secondary structure. The JNet
\r
11 method uses several different neural networks and decides on the
\r
12 most likely prediction via a jury network. <br>
\r
14 <li>Cuff J. A and Barton G.J (1999) Application of enhanced multiple
\r
15 sequence alignment profiles to improve protein secondary
\r
16 structure prediction <em>Proteins</em> <strong>40</strong> 502-511</li></ul>
\r
18 The function available from the <strong>Web Service→Secondary
\r
19 Structure Prediction→JNet Secondary Structure
\r
20 Prediction</strong> menu does two different kinds of prediction,
\r
21 dependent upon the currently selected region:</p>
\r
23 <li>If nothing is selected, and the displayed sequences appear to
\r
24 be aligned, then a JNet prediction will be run for the first
\r
25 sequence in the alignment, using the current
\r
26 alignment. Otherwise the first sequence will be submitted for prediction.
\r
29 just one sequence (or a region on one sequence) has been selected,
\r
30 it will be submitted to the automatic JNet prediction server
\r
31 for homolog detection and prediction.
\r
33 <li>If a set of sequences are selected, and they appear to be aligned,
\r
34 then the alignment will be used for a Jnet prediction on the
\r
35 <strong>first</strong> sequence selected in the set (that is, the one
\r
36 nearest the top of the alignment window).
\r
39 <p>The result of a JNet prediction for a sequence is a new annotated
\r
40 alignment window:</p>
\r
41 <img src="jnetprediction.gif">
\r
43 The sequence for which the prediction was made is the first one in the
\r
44 alignment. If a sequence based prediction was made then the remaining
\r
45 sequences in the alignment are the aligned parts of homologs which
\r
46 were used to construct a sequence profile for the prediction. If the
\r
47 prediction was made using a multiple alignment, then the original
\r
48 multiple alignment will be returned, annotated with the prediction.
\r
50 The annotation bars below the alignment are as follows:</p>
\r
52 <li>Lupas_21, Lupas_14, Lupas_28<br>
\r
53 <em>Coiled-coil predictions for the sequence. These are binary
\r
54 predictions for each location.</em></li>
\r
55 <li>JNETSOL25,JNETSOL5,JNETSOL0<br>
\r
56 <em>Solvent accessibility predictions - binary predictions of 25%, 5%
\r
57 or 0% solvent accessibility.</em></li>
\r
59 <em>The consensus prediction - helices are marked as red tubes, and
\r
60 sheets as dark green arrows.</em></li>
\r
62 <em>The confidence estimate for the prediction. High values mean high
\r
63 confidence. prediction - helices are marked as red tubes, and
\r
64 sheets as dark green arrows.</em></li>
\r
66 <em>Alignment based prediction - helices are marked as red tubes, and
\r
67 sheets as dark green arrows.</em></li>
\r
69 <em>HMM profile based prediction - helices are marked as red tubes, and
\r
70 sheets as dark green arrows.</em></li>
\r
72 <em>Jpred prediction - helices are marked as red tubes, and
\r
73 sheets as dark green arrows.</em></li>
\r
75 <em>PSSM based prediction - helices are marked as red tubes, and
\r
76 sheets as dark green arrows.</em></li>
\r
78 <em>Amino Acid frequency based prediction - helices are marked as red tubes, and
\r
79 sheets as dark green arrows.</em></li>
\r
81 <em>A '*' in this annotation indicates that the JNETJURY was invoked
\r
82 to rationalise significantly different primary predictions.</em></li>
\r