Information Technology Reference
In-Depth Information
5 Meta-Servers for Protein Tertiary Structure Prediction
Several meta-servers not only integrate protein structure predictions performed by
various methods but also assemble and interpret the results to come up with a
consensus prediction. This section deals with a comprehensive discussion of such
meta-servers.
Pcons.net meta-server (Wallner et al.
2007
) retrieves results from several pub-
lically available servers which are then analyzed and assessed for structural cor-
rectness using Pcons as well as ProQ, thus presenting the users a ranked list of
possible models (Lundstr
m et al.
2001
). In combination of several publically
available servers, Pcons.net meta-server also uses Reversed Position Speci
รถ
c
BLAST (RPS-BLAST) to parse the sequence into structural domains by analyzing
the signi
cance and span of the best RPS-BLAST alignment.
3D-Jury (Ginalski et al.
2003
) are the meta-servers which focus on the selection
of high quality obtained from different servers. 3D-Jury, takes groups of models
generated by a set of servers as input which are then compared with each other and
a similarity score is assigned to each pair by MaxSub tool (Siew et al.
2000
)
followed by providing ranking to the models.
3D-SHOTGUN (Fischer
2003
) meta server does not just select the best model
but also re
nes initial models for building the protein structure model with high
accuracy. 3D-SHOTGUN meta-predictor consists of three steps: (i) assembly of
hybrid models, (ii) con
rst
assembles hybrid models from the initial models and then assigns scores to each of
the assembled models by using the original models scores and the structural sim-
ilarities between them. Thereby resulting a highly sensitive and ensuring a signif-
icantly higher speci
dence assignment, and (iii) selection. 3D-SHOTGUN
city of the models than that of individual servers (Fischer
2003
).
GeneSilico (Kurowski and Bujnicki
2003
) is another meta-server which com-
bines the useful features of other meta-servers available, but with much greater
flexibility of the input in terms of user-defined multiple sequence alignments.
However, there are several drawbacks reported in the current meta-servers
including 3D-Jury (Ginalski et al.
2003
) and GeneSilico (Kurowski and Bujnicki
2003
). They take the initial threading inputs from remote computer which are
occasionally shut down or are not available. Secondly, the instability of the algo-
rithms of the remote servers is another drawback of these meta-servers (Wu and
Zhang
2007
).
LOMETS (Wu and Zhang
2007
), overcomes the above drawbacks. It is one of
the good performing meta-servers in which all nine individual threading servers are
installed locally, which facilitates controlling and tuning of Meta-server algorithms
in a consistent manner making the users able to obtain quick
final consensus. It
facilitates quick generation of initial threading alignments owing to the nine state of
art threading programs that are installed and run in a local computer cluster, thus
ensure faster results as compared to the traditional remote-server-based meta-
servers. Based on TM-score,
the consensus models generated from the top
Search WWH ::
Custom Search