1. Summary
 | |
TOBMODEL server takes amino acid sequence as input and generates c-alpha 3D models of the putative Transmembrane beta barrel protein. Topologies are generated using BOCTOPUS and refer to the number of beta strands and in and out loop location. Currently the top ranked model is chosen based on the minimum error between the predicted Z-coordinate and Z-coordinates obtained from a 3D model. In future, more complex methods can be employed to select the top ranked 3D model.
Note that the server does not predict cleavable signal peptides, which are easily confused with TM segments. If signal peptides are likely to be present in the input data, a separate signal peptide predictor such as SignalP should first be applied and predicted signal peptides cleaved off before submitting the sequence to TOBMODEL.
| |
2. Usage
| |
This server is appropriate for 3D model generation of putative transmembrane beta barrel proteins. Input to the server is amino acid sequences in FASTA format. The current version of TOBMODEL is ideal for 1 protein sequence per input. If more than 1 input sequences are input then only the first sequence is processed. TOBMODEL obtains predicted topologies from BOCTOPUS. We are currently working on a faster version of BOCTOPUS for multi-sequence input.
Example input:
>1qj8A:A|PDBID|CHAIN|SEQUENCE
ATSTVTGGYAQSDAQGQMNKMGGFNLKYRYEEDNSPLGVIGSFTYTEKSRTASSGDYNKN
QYYGITAGPAYRINDWASIYGVVGVGYGKFQTTEYPTYKNDTSDYGFSYGAGLQFNPMEN
VALDFSYEQSRIRSVDVGTWIAGVGYRF
>1TLY:B|PDBID|CHAIN|SEQUENCE
AENDKPQYLSDWWHQSVNVVGSYHTRFGPQIRNDTYLEYEAFAKKDWFDFYGYADAPVFF
GGNSDAKGIWNHGSPLFMEIEPRFSIDKLTNTDLSFGPFKEWYFANNYIYDMGRNKDGRQ
STWYMGLGTDIDTGLPMSLSMNVYAKYQWQNYGAANENEWDGYRFKIKYFVPITDLWGGQ
LSYIGFTNFDWGSDLGDDSGNAINGIKTRTNNSIASSHILALNYDHWHYSVVARYWHDGG
QWNDDAELNFGNGNFNVRSTGWGGYLVVGYNFHHHHHH
| |
3. Output
| |
The server generates a C-alpha PDB file as an output. It also outputs prediction results in one text file and generates a graphical representation of the predicted topologies and svm probabilitiess:
C-alpha PDB file (tmbmodel_ca.pdb): PDB file containing the top ranked c-alpha 3D model.
Representative topologies obtained from topology clusters (tmbmodel_topology_cluster.txt): Representative topologies from predicted topology clusters, the number of strands, frequency of occurance of that topology and the average Zerr per strand for each cluster.
Predicted topologies (topologies.txt): Text file containing the predicted topologies for the input protein, where i=inside (cytosolic side), M=membrane and o=outside (extra-cellular side).
Graphical output (topologies.png): A graphical representation of predicted topology and svm probabilities. For representation purpose of the overall topology, M is assigned a value 2, i and o are assigned 1.8 and 2.2, respectively. Green, gray and blue lines represent the per-residue preferance to be in M,i,o.
Rank file (tmbmodel_rank.txt): Text file containing the top ranked topologies, their shear number, Zerr and predicted number of strands.
Input sequence (inputsequence.txt): input amino acid sequence.
Profile (profile.txt): Profile comprising of SVM outputs that is then used as an input to the HMM.
PSSM (pssm.txt): Position specific scoring matrix generated for the input sequence.
| |
4. References
| |
BOCTOPUS:
BOCTOPUS: Improved topology prediction of transmembrane beta barrel proteins. Sikander Hayat and Arne Elofsson (2012)Bioinformatics 28(4), 516-522. [Pubmed]
OCTOPUS:
OCTOPUS: improving topology prediction by two-track ANN-based preference scores and an extended topological grammar. Håkan Viklund and Arne Elofsson (2008) Bioinformatics 24(15), 1662-1668. [Pubmed]
TOBMODEL:
TOBMODEL: Toward 3D modeling of transmembrane beta barrel proteins based on z-coordinate and topology prediction. Sikander Hayat and Arne Elofsson (ISMB 2012)
Full Atom model based on:
Gront, D., Kmiecik, S., and Kolinski, A. (2007). Backbone building from
quadrilaterals: A fast and accurate algorithm for protein backbone reconstruction
from alpha carbon coordinates. Journal of computational chemistry, 28(9), 1593-1597.
| |
5. Contact
|
|
