possible to view the results in html or xml format by clicking
the corresponding boxes on the upper right-hand corner and
to export the graphical output to an image. Results can be
reached any time from “My Predictions” button after logging
in to the system.There are also stand alone tools such as INTREPID that can be
used by downloading and installing their packages, instead of doing
interactive analysisviaweb servers. Users having advanced knowl-
edge in scripting may refer to these tools and run them on their
local computers.3 Case Study: Human KIT Protein and Its Predicted Functional Sites
As an example, human KIT protein (UniProt accession: P10721) is
chosen to illustrate the functional site prediction procedure using
the defined work-flow (including TraceSuite II), Consurf, Univer-
sal Evolutionary Trace and PROFisis. KIT protein, which is a kinase
inhibitor, has a crystal structure in a bound state with imatinib (i.e.,
an anticancer drug) in PDB (id: 1t46). Using this structure and
KIT’s sequence, we aimed to find the residues in contact with the
drug molecule (i.e., functional sites). The residues having contact
with imatinib are experimentally known; however, we have only
used this information in order to compare the prediction results
with the known sites.
At the beginning of the analysis we have obtained the sequence
of the KIT protein from the UniProt database and assumed that
this is an unknown sequence. Then, we searched for the homologs
of our query using KIT protein’s sequence via UniProt BLAST tool
with default parameters. The top ranked result of the BLAST
analysis was naturally the actual KIT protein record in UniProt.
After that, the homologous sequences found at the BLAST analysis
(except the hits with 100% identity to the query and the fragment
sequences) are analyzed with MSA using the Align tool in UniProt.
The output MSA was investigated manually by highlighting the
known functional sites on the aligned sequences. Following MSA
analysis, fasta formatted output MSA was given to TraceSuite II
together with the structural information (by changing the FASTA
header of the query sequence to “>1t46” in the MSA output). The
results of TraceSuite II (i.e., the evolutionary trace mapped on
sequences and on the given structure, and the constructed phylo-
genetic tree) have been investigated. Finally, human KIT protein
was analyzed with ConSurf, Universal Evolutionary Trace, and
PROFisis, by directly entering its sequence to the corresponding
web-services. In ConSurf analysis, the structural information was
not given at the input.64 Heval Atas et al.