structure (only available if the structural information is given at the
input level), for each partition with highlighted buried/exposed
residues (seeNote 4). Clicking the given PDB id at the bottom of
the results page will reveal the evolutionary trace mappings for
different partitions on the corresponding structure. Residue posi-
tions in the output needs to be checked to match with the positions
in the query sequencePQ(seeNote 5). However, it is easier to
follow the positions of the annotated residues on the structures. At
the “evolutionary trace mapped to each structure” page, clicking
on “Molscript input file” will display the list of buried/exposed
residues with their respective positions on the structure. One
important note here is that, the annotated residue positions on
the structure and the query sequence usually does not match
since the crystal structures are only available for some parts of
protein sequences. It is possible to observe the position of a residue
on the crystallized structure (as opposed to the position on the
whole protein sequence), using the PDB service (e.g., residues
position information for the crystal structure of KIT protein:
https://www.rcsb.org/pdb/explore/remediatedSequence.do?
structureId¼1T46).
Partitions illustrated in the output are referred from the con-
structed phylogenetic tree on which each partition is labelled with a
vertical line. Prediction is more stringent in the first partition, as
there are more classes of homologs compared to the second parti-
tion. From the first partition to the last partition, more residues are
labelled as important. In TraceSuite II, residues vary among distant
homologs (homologs across the classes obtained from the parti-
tioning) rank better than the residues vary among close homologs
(homologs within the same class) in the phylogenetic tree. At the
evolutionary trace output on the main results page, the buried
residues are marked according the tenth partition, which is the
main interest.
Up to this point, we have demonstrated a workflow to infer
functional sites on a query sequence. Below, the alternative ways to
obtain similar information using all-in-one tools, are explained.2.5 All-in-One Web
Servers for Functional
Site Prediction
(a) Universal Evolutionary Trace (ET)
Universal ET tool works with the same principle as Trace-
Suite II; however, it either takes an amino acid sequence in
FASTA format, a UniProt accession, a PDB identifier or a
PDB structure file as an input, and calculates the functionally
important residues following a methodology similar to the
one described above. Universal ET stores the precomputed
evolutionary traces for UniProt protein records and PDB
structure records; as a result, when queried with accessions/
ids the system is able to display the precomputed results
without any computation, which accelerates the analysis.
Entering protein structure information at the input level60 Heval Atas et al.