DictyOGlyc - 1.1

Submission

---

Sequence submission: paste the sequence(s) or upload a local file

Paste a single sequence or several sequences in FASTA format into the field below:

Submit a file in FASTA format directly from your local disk:

Generate graphics   

---

Restrictions:
At most 50 sequences and 70,000 amino acids per submission; each sequence not more than 4,000 amino acids.
Note that the first 10 characters (after '>') in the fasta header line must be unique.

Confidentiality:
The sequences are kept confidential and will be deleted after processing.

---

CITATIONS

For publication of results, please cite:

Scanning the available Dictyostelium discoideum proteome for O-linked GlcNAc glycosylation sites using neural networks.
R. Gupta, E. Jung, A.A. Gooley, K.L. Williams, S. Brunak and J. Hansen.
Glycobiology: 9(10):1009-22, 1999.

View the abstract.

---

O-GlycBase

See the O-GlycBase database for a revised database of O- and C-glycosylated proteins.

Instructions

---

In order to use the DictyOglyc server for prediction on amino-acids sequences either: 

1. Type the name of the sequence in the 'Sequence name' field 
• The sequence must be submitted using the one letter code for the amino acids:
  `acdefghiklmnpqrstvwyACDEFGHIKLMNPQRSTVWY' 
• Other characters will be accepted, but not encoded in the network window, when making the prediction. Do not use these. 
• Peptides shorter than 19 residues can produce non-reliable results. (always include 9 residues on both sides of the Ser/Thr you want evaluated. The optimal is to use the complete "native" glycoprotein sequence truncating the signal sequence)
2. Click on the "Generate Graph" checkbox if you'd like to see a graph with the output (Gif/Postscript). These may be quite useful in seeing the "hot" spots in your protein. 
3. Press the "Submit sequence" button. 
4. A WWW page will return the results when the prediction is ready. Response time depends on system load. 

OR

5. Use the file browser to find your sequences in FASTA format in your own appropiate directory (a single file may contain multiple sequences).

Example of two sequences in fasta format:

>seq1
ASTPGHTIIYEAVCLHNDRTTIP
>seq2
ASQKRPSQRHGSKYLATASTMDHARHGFLPRHRDTGILDSIGRFFGGDRGAPK
NMYKDSHHPARTAHYGSLPQKSHGRTQDENPVVHFFKNIVTPRTPPPSQGKGR
KSAHKGFKGVDAQGTLSKIFKLGGRDSRSGSPMARR

6. Click on Generate graph if required, and then press the "Send file" button. 

Output format

---

If the Potential > Threshold then O-glycosylation is predicted at that site. Predicted sites are assigned 'G'.
In this example Thr 13, 17, 19 and 21 are predicted to be O-glycosylated.
The higher the potential the higher is the confidence of the prediction.
The Threshold takes into account surface/buried predictions of the particular site, and indicates
a value above which sites are predicted to be on the surface and O-glycosylated.
The figure (using: Generate graph) shows the potential and threshold versus sequence position.
Alternative non-predicted sites may be found where the potential is high (>0.4) and threshold low (<0.55).

Name:  test_seq         Length:  22

GSDWSGVCKITATPAPTVTPTV <-- Submitted sequence.

............G...G.G.G. <-- The assignments. G stands for predicted

                           GlcNAc O-glycosylation at site.



SeqName    Residue   Number  Potential Threshold   Assignment

test_seq     Ser     0002    0.3727     0.4493         . 

test_seq     Ser     0005    0.3553     0.4687         . 

test_seq     Thr     0011    0.3973     0.4687         . 

test_seq     Thr     0013    0.5762     0.4606         G 

test_seq     Thr     0017    0.5945     0.4567         G 

test_seq     Thr     0019    0.6180     0.4638         G 

test_seq     Thr     0021    0.5827     0.4455         G

Graphics in PostScript

Changes in v/1.1

---

The training dataset has been slightly enlarged to include more 'negative' (non-glycosylated) sites.

Surface-derived threshold has been modified to perform a running average.

Glycosylation potentials have been sigmoidally scaled to magnify the distinction between positive and negative sites.

We expect the server to predict sites more 'specifically', that is to pick upO-GlcNAc sites fromsecreted and membrane proteins of D.discoideum. Predictions with the new server may vary from those made with DictyOGlyc v/1.0. On a general basis, the server picks fewer glycosylation sites presumably excluding more false positives.

Feedback, Comments and suggestions are most welcome.