The ConSurf Server colors amino acids in a 3D protein structure
by evolutionary conservation -- entirely automatically.
For details, please see the
Beginning with ConSurf 3,
the results of a ConSurf job can be saved in a specially annotated PDB file.
The examples below were saved from ConSurf jobs. The parameters of each job
can be viewed (after you display the result using one of the links
below) by clicking ConSurf Job Information.
2VAA chain A:
Major histocompatibility class I (mouse) complexed with an octapeptide from
Key points to appreciate about this intriguing pattern of conservation are
noted in views saved from the ConSurf result as
2VAA chain A: Insufficient data.
150 sequence homologs were used for the above job, but in this job,
only 50 were used. With only 50 sequences, conservation grades could
not be calculated with confidence for all amino acids. A special
yellow color indicates the amino acids for which the data were
insufficient to calculate a reliable conservation grade. Detecting
and marking amino acids with insufficient data is a new feature in
ConSurf 3. After you display this result, you will see a link "Display #1"
that enables you to hide the insufficient data color if you wish.
2VAA chain A: Alternate color scheme.
This illustrates how the default conservation color scheme used by
ConSurf can be changed in a saved PDB file. (In this
example, grades 4, 5, and 6 use the same color.)
The default colors work best on a white background, but here, the
default background has been changed to black, which works best for
the alternate scheme.
Instructions for changing the conservation color scheme are available
by clicking the link "Changing the conservation color scheme" (at the bottom
of the ConSurf
or by clicking
1FLO chain A: Flp recombinase complexed with a DNA Holliday junction.
Any combination of the four sequence-identical chains in this homotetramer
can be colored according to the ConSurf results (see the checkboxes for
each chain). This is a new feature in ConSurf 3. There are many amino acids
with insufficient data here because only nine homologs were found
in Uniprot. However, the data are sufficient to see
conservation in the DNA-binding and catalytic regions with confidence.
1IGY chain B:
Heavy chain of an intact IgG1 antibody.
Illustrates a case in which ConSurf's automated gathering of homologous
sequences leaves one or more entire
domains with insufficient data for
reliable assignment of conservation grades.
Which domains have insufficient data varies with time, as the sequence databases
grow. (When this example was run, in June 2006, the entire variable domain
had insufficient data.)
Whole domains with insufficient data
is a common complication with multiple domain
chains. To avoid this, it is recommended that you separate the single domain
of interest from the PDB file (by text editing), and upload the
(The absence of adequate data for the constant domains was not
evident in ConSurf version 2, but could be detected when you
inspected the multiple sequence alignment carefully.)
Variable Domain of Heavy Chain:
In order to guarantee that conservation grades are calculated for the
variable domain of the antibody heavy chain, a separate run was performed
by uploading a PDB file in which the constant domains (and corresponding
SEQRES segment) were deleted for chain B (keeping only residues 2-114
in chain B). (The SEQRES must be changed since ConSurf uses it!)
The striking patch of highly variable residues occurs at the antigen-binding
face (complementarity-determining regions). The highly conserved residues
appear to be those crucial for proper folding of this immunoglobulin
This is also a case
in which the sequence contains "insertions"
because the sequence is numbered according to
a reference sequence. There are also many "gaps" in the sequence numbering
for the same reason. In this case, these do not represent physical gaps in the
To see the inserted blocks, click on Clickable ConSurf-Colored Sequence 3D;
they are marked with green asterisks.
Such insertions were not handled correctly
by ConSurf version 2.
Seed protein crambin (single unnamed chain).
This PDB file contains sequence microheterogeneity, which
was not handled correctly by ConSurf 2, but is handled better by
Sequence microheterogeneity can best be detected by a
careful inspection of the contents of the PDB file. Typically, it is
signified by the occurrence
of residues with coordinates (ATOM records) that are
not listed in the SEQRES records. Typically, in the coordinates,
these have redundant sequence numbers with insertion letters. For
example, in 1CBN, there is microheterogeneity at positions
22 and 25, noted in REMARK 4. This causes the number of residues
in SEQRES to be less than the number with coordinates (shown in the
ConSurf Job Information report). When this occurs, Protein Explorer
asks you to confirm that the PDB file contains sequence microheterogeneity,
as otherwise, this condition might signal an internal error in the ConSurf
Sequence microheterogeneity typically shows up as "insertions"
in Protein Explorer's ConSurf-Colored Sequence 3D, Seq3D,
and Sequences listings, because Protein Explorer cannot presently
distinguish true insertions (present in both SEQRES and coordinates)
from sequence microheterogeneity.
ConSurf calculates conservation grades for the residues
in the SEQRES record, which do not include the "inserted"
residues representing microheterogeneity. Therefore, the latter
are not assigned conservation colors, and remain colored carbon gray.
a list of PDB files that illustrates a range of technical issues for ConSurf
and Protein Explorer.
Rare PDB files are not handled correctly by PE, and therefore their ConSurf
results are also not handled correctly by PE.
Feedback to Eric Martz.