(if available) will improve the functional importance inference
in most cases.
SearchPQin Universal ET by first selecting “Amino acid
sequence in FASTA format” and entering its sequence to the
corresponding field and click Run trace button (http://mam
moth.bcm.tmc.edu/uet/). A progress page will be displayed
and refreshed every 10 s, which includes information on each
step of the procedure. After the run is finished, the results page
displays the query sequence in a way that each residue is
colored according the functional importance, using the color
range in the visible light spectrum (i.e., red: more important,
green: mid-range, and blue/purple: less important). It is also
possible to visualize the ET ranks as histograms by clicking the
corresponding button (Fig.3). The results page also displays
the generated phylogenetic tree in circular or rectangular
layout. It is also possible to download all of the intermediate
and resulting files in a compressed file. Here, the residues that
are predicted to be important are the inferred functional sites
forPQ. The user may move on to the Subheading2.5,item b
for more predictions.
(b) ConSurf
The second alternative method, ConSurf, accepts either a
sequence or an MSA or a phylogenetic tree as input. As a
result, ConSurf can be employed at any step of the workflow
described above. Below, the methodology is described for
sequence queries.
QueryPQvia ConSurf web server (http://consurf.tau.ac.
il/2016/) by entering its sequence and selecting “Amino-
Acids” (it is also possible to analyze nucleotides) as the
sequence type. If there is a known protein structure in PDB
for the query sequence, it can be added in the next step either
by entering PDB id or uploading the PDB file. If the query is a
protein sequence without a known 3D structure, it can be
predicted via MODELLER in the presence of a license key
(optional). In the following step, paste the query sequence
into the box in FASTA format for ConSurf to run MSA (the
sequence will be extracted directly from PDB file if 3D
äFig. 3Human KIT protein and its predicted functional sites. The selected region of KIT protein involving
imatinib binding sites (residues between 594 and 812) is represented in three blocks and the results of four
computational methods are aligned with the query sequence, on the horizontal axis. Blue colored residues on
the query sequence shows the experimentally known imatinib binding sites. The predicted buried and
conserved positions (i.e., functional sites) are highlighted with light gray in TraceSuite II output. Universal
ET outputs a histogram colored according to the importance of the residues, where the importance increases
from blue to red. ConSurf colors residues based on their conservation scores, as well. It also labels buried/
exposed and functional/structural residues. PROFisis displays the predicted protein binding sites, secondary
structures and solvent accessibilities on the query sequence
62 Heval Atas et al.