A combined transmembrane topology and signal peptide predictor
Phobius takes proteins in FASTA format. It recognises the 20 amino acids plus B, Z and X, which are all treated as unknown. Any other character is removed. Lowercase input is accepted and converted to uppercase.
This is an example (one protein):
>Q8TCT8|PSL2_HUMAN you can have comments after the ID MGPQRRLSPAGAALLWGFLLQLTAAQEAILHASGNGTTKDYCMLYNPYWTALPSTLENAT SISLMNLTSTPLCNLSDIPPVGIKSKAVVVPWGSCHFLEKARIAQKGGAEAMLVVNNSVL FPPSGNRSEFPDVKILIAFISYKDFRDMNQTLGDNITVKMYSPSWPNFDYTMVVIFVIAV FTVALGGYWSGLVELENLKAVTTEDREMRKKKEEYLTFSPLTVVIFVVICCVMMVLLYFF YKWLVYVMIAIFCIASAMSLYNCLAALIHKIPYGQCTIACRGKNMEVRLIFLSGLCIAVA VVWAVFRNEDRWAWILQDILGIAFCLNLIKTLKLPNFKSCVILLGLLLLYDVFFVFITPF ITKNGESIMVELAAGPFGNNEKLPVVIRVPKLIYFSVMSVCLMPVSILGFGDIIVPGLLI AYCRRFDVQTGSSYIYYVSSTVAYAIGMILTFVVLVLMKKGQPALLYLVPCTLITASVVA WRRKEMKKFWKGNSYQMMDHLDCATNEENPVISGEQIVQQ
This server accepts up to 100 sequences and 50,000 residues per request; a single sequence may be at most 10,000 residues. For whole proteomes, install the standalone package.
The long format lists the predicted transmembrane helices, the intervening loop regions and the signal peptide, as a UniProt-style feature table:
ID MTH_DROMEa signal peptide FT SIGNAL 1 24 FT REGION 1 3 N-REGION. FT REGION 4 19 H-REGION. FT REGION 20 24 C-REGION. FT TOPO_DOM 25 218 NON CYTOPLASMIC. FT TRANSMEM 219 238 FT TOPO_DOM 239 249 CYTOPLASMIC. //
If the whole sequence is labelled cytoplasmic or non-cytoplasmic, the prediction is that it contains no membrane helices. It is not wise to interpret that as a prediction of location. The prediction gives the most probable location and orientation of transmembrane helices, found by an algorithm called N-best (1-best in this case) that sums over all paths through the model with the same location and direction of helices.
One line per protein: the identifier, then the number of predicted transmembrane segments, a Y/0 indicator for a signal peptide, and the topology.
SEQENCE ID TM SP PREDICTION MTH_DROME 7 Y n4-19c24/25o219-238i250-269o281-302i
Helix positions are separated by i where the loop is cytoplasmic
and o where it is non-cytoplasmic. A signal peptide is given as the
position of its h-region between n and c, followed by
the last residue of the signal peptide and the first of the mature protein,
separated by /.
The plot shows the posterior probability that each residue is cytoplasmic, non-cytoplasmic, part of a transmembrane helix, or part of a signal peptide. Weak helices that did not make it into the prediction show up here, as does the certainty of each segment. The bar under the axis shows the one-best prediction.
Plot and prediction can look contradictory, because the plot shows per-residue probabilities whereas the prediction is the single most probable overall structure. Treat the plot as complementary information.
If you already know where part of the protein sits, fix it on the constrained prediction page and Phobius will find the most probable topology consistent with your constraints.
Topology is more conserved than sequence, so PolyPhobius can use an alignment of homologues to sharpen a prediction. Supply an aligned FASTA file—the prediction is reported for the first sequence. Alternatively, submit a single sequence and the server will search Swiss-Prot and build the alignment for you.
See the API page for a JSON endpoint suitable for scripting.
Please report problems via the issue tracker.