Secondary and Tertiary Structure Prediction of Proteins: A Bioinformatic Approach - Complex System Modelling and Control Through Intelligent Soft Computations - page 547

solvation terms and finally, evaluation of the alignment using a neural network

(Jones 1999b ). GenTHREADER is advantageous as apart from being very fast, it
requires no human intervention in the prediction process.
FUGUE (Shi et al. 2001 ) is another example of hybrid server for recognizing
distant homologues by sequence-structure comparison. FUGUE utilizes environ-
ment-speci c substitution tables and structure-dependent gap penalties. Here scores

for amino acid matching and insertions/deletions are evaluated based on the local
structural environment of each amino acid residue in a known structure. Local
structural environment de

ned in terms of secondary structure, solvent accessibil-
ity, and hydrogen bonding status, are used by FUGUE to produce a high quality 3D
protein model. FUGUE also encompasses scanning database of structural pro les,

calculation of the sequence-structure compatibility scores and prediction of align-
ment of multiple sequences against multiple structures in order to enrich the con-
servation/variation information (Shi et al. 2001 ).
123D+ ( http://pole-modelisation.univ-bpclermont.fr/prive/

ches_HTML/123D
+.html ) is another hybrid server which combines sequence pro les, secondary

structure prediction and contact capacity potential to thread a protein sequence
through asset of structures.
RaptorX ( http://raptorx.uchicago.edu/ ) is a protein structure prediction hybrid
server that excels in predicting 3D structures for protein sequences without close
homologs in the PDB (K ä llberg et al. 2012 ). It predicts secondary and tertiary
structures, contacts, solvent accessibility, disordered regions and binding sites for a
given input sequence. Raptor X,

le-entropy scoring method to
assess the quality of information content in sequence pro

first of all uses pro

les (Peng and Xu 2010 ).

Thereafter it uses conditional random fields to integrate a variety of biological

signals in a nonlinear threading score. Finally, multiple-template threading proce-
dure (Peng and Xu 2009 ), which enables the use of multiple templates to model a
single target sequence is used to produce a high quality protein 3D model.
MULTICOM toolbox ( http://sysbio.rnet.missouri.edu/multicom_toolbox/ ) s
another programme consisting of a set of protein structure and structural feature
prediction tools. Secondary structure prediction, solvent accessibility prediction,
disorder region prediction, domain boundary prediction, contact map prediction,
disul

de bond prediction, beta-sheet topology prediction, fold recognition, multiple
template combination and alignment, template-based tertiary structure modeling,
protein model quality assessment, and mutation stability prediction are some of the
functions facilitated by MULTICOM toolbox (Cheng et al. 2012 ).
Hybrid methods use various aspects for predicting an accurate protein tertiary
structure. However Meta-servers discussed in the following section deals with
generation of a consensus prediction of protein tertiary structure assembled from
different servers.