SVMtop Server 1.0 Help

This server predicts transmembrane (TM) helices and their topologies in membrane proteins. The first predicted TM helix within 60 residues of the
N-terminus is sometimes a signal peptide. If your proteins contain signal peptides, a signal peptide prediction program would be better suited
for identifying their locations. Do NOT use this program to predict if the protein is cytoplasmic or secreted.

TABLE OF CONTENT

INPUT
HOW TO SUBMIT
HELIX PREDICTION SETTINGS
TOPOLOGY PREDICTION SETTINGS
OUTPUT

INPUT

Please input the protein sequence in FASTA format. Single-letter representation of 20 amino acids is accepted (excluding B, X, and Z).
Please check your sequence before submitting to the program. The word following the character “>” is the identifier of the sequence.
The sequence ends if another line starting with a ”>” appears; this indicates the start of another sequence.

Here is an example:

>ROM1_BOVIN
MAPVLPLVLPLQPRIRLVQGLWLLSWLLVLVGGLTLLCSGHLLVQLWHLGTFLAPSCPFSALPQVALAASAVALGTGLVGSGASRASLDAEQYPPWRGVLGPLLVAGTAGGGGLLVLALGLALALPGTLDTGLEEGLGSALVHYKDTEVPGRCQAKRLLDELQLRHHCCGRHGYKDWFGIQWVSNRYLDPNDPDVVDRIQSNVEGLYLIDGVPFSCCNPHSPRPCLQSQLSDPHAHPLFDPRQPNLNLWSQGCHEVLLGHLQGLASTLGNMLAVTFLLQTLVLLGLRYLQTALEGLGGVIDGEGEAQGYLFPAGLKDMLKTAWLQGAGPHRPAPGETPPEEKPPKECLPEA
>COXK_BOVIN
FENRVAEKQKLFQEDNGLPVHLKGGATDNILYRVTMTLCLGGTLYSLYCLGWASFPHKK

HOW TO SUBMIT

To use the basic setting (default), click "Submit" once you have entered the sequence(s) into the box.
If you would like to use the advanced setting, here are some explanations:

HELIX PREDICTION SETTINGS

l_min: the minimum length of predicted helix segment; if a predicted helix segment is less than the l_min, it will be predicted as non-helix (default = 7).
l_max: the maximum length of predicted helix segment; if a predicted helix segment is greater than l_max, it will be split into two helices in the center
        with insertion of l_gap non-helix residues (default = 33).
l_opt: the optimum length of helix to be extended if the predicted TM segment is greater than l_min and less than the defined l_opt (default = 21).
l_gap: the number of residues to be inserted in the center of the split TM segment (default = 2).

If you would like to change the helix prediction settings, we suggest modifying l_min and l_max first.

TOPOLOGY PREDICTION SETTINGS

b: the base parameter in the Alternating Geometric Scoring Function (AGSF); suggested range (1.2 ≤ b ≤ 3.0) (default = 1.6)
EI: the exponent increment in the AGSF; suggested range (0.5 ≤ EI ≤ 3.0) (default = 1.0)

OUTPUT

Short Output Description

ROM1_BOVIN Protein length: 351
ROM1_BOVIN Number of predicted TMH: 4
ROM1_BOVIN Predicted topology of N-term: Inside

Some explanations of the output:

Ÿ Protein length: the total length of the query protein
Ÿ Number of predicted TMH: total number of predicted TM segments
Ÿ Predicted topology of N-term: The predicted topology of the N-terminal loop

Detailed output description

ROM1_BOVIN Protein length: 351
ROM1_BOVIN Number of predicted TMH: 4
ROM1_BOVIN Predicted topology of N-term: Inside

ROM1_BOVIN TS_i = 1.2348; TS_o = 0.3424
Probability of predicted N-term topology: 0.7829
#

Some explanations of the output:

Ÿ The first line under the sequence of the query protein is the predicted label of each residue (i: inside loop; H: helix; o: outside loop).
Ÿ The second line under the sequence of the query protein is the probability of each residue carrying the predicted label.
Ÿ TS_i = Topology score of N-term loop as inside; TS_o = Topology score of N-term loop as outside
Ÿ Probability of predicted N-term topology: the probability of the predicted N-term topology as inside or outside.