Secondary and Tertiary Structure Prediction of Proteins: A Bioinformatic Approach - Complex System Modelling and Control Through Intelligent Soft Computations - page 539

Table 3 List of web servers for modeling protein structure by ab initio folding method along with
the webpage URL and the programme description
S. no. Name of the web server/group

(URL)

Description of the web server/group

1 ROBETTA (Kim et al. 2004 ;

Bradley et al. 2005 )[ http://
robetta.bakerlab.org ]

This web-server provides ab initio and
comparative models of protein domains.
Domains having no sequence similarity
with PDB sequences are modeled by
Rosetta de novo protocol

2 QUARK (Xu and Zhang 2012 )

[ http://zhanglab.ccmb.med.
umich.edu/QUARK/ ]

De novo protein structure prediction web server
aims to construct the correct protein 3D model
from amino acid sequence by replica-exchange
Monte Carlo simulation under the guide of an
atomic-level knowledge-based force eld

3 PROTINFO (Hung et al. 2005 )

[ http://protinfo.compbio.
washington.edu ]

De novo protein structure prediction web
server utilizes simulated annealing for 3D
structure generation and different scoring
functions for selection of nal five conformers

4 SCRATCH (Cheng et al. 2005 )

[ http://www.igb.uci.edu/servers/
psss.html ]

This server utilizes recursive neural networks,
evolutionary information, fragment libraries and
energy to build protein 3D model

5 BHAGEERATH (Jayaram et al.

2006 )[ http://www.scfbio-iitd.
res.in/bhageerath ]

Energy based methodology for narrowing down
the search space and thus helps in building a
good protein 3D model

completely predicts a new fold (Skolnick and Kolinski 2002 ; Floudas et al. 2006 ).
With increasing protein size, the conformational space to be searched increases
sharply, this makes the ab initio modeling of larger proteins extremely dif cult

(Zhang and Skolnick 2004 ).
Currently, the accuracy of ab initio modeling is limited to small proteins having
length less than 50 amino acid residues. Ab initio structure prediction requires an
ef find the conformation of the modeled protein near to

native state protein structure with lowest free energy. Ab initio structure prediction
is challenging because current potential functions have limited accuracy. Few
popular web servers for modeling of the protein structure by ab initio folding/
modeling method are listed in Table 3 .

cient potential function to

3.2 Fold Recognition (FR) or Threading

Fold recognition or threading method aims to fit a target sequence to a known
structure in a library of folds and the model built is evaluated using residue based
contact potentials (Floudas 2007 ). Although fold recognition will not yield equiv-
alent results as those from X-ray crystallography or NMR yet, it is a comparatively