|1.48 ||24/10/2013 || ||
- Error in web page: 'Non-existent PDB entry'.
- Error in XML for assembly list.
- Some assemblies not displayed correctly.
|1.47 ||21/03/2013 || ||
- Database search results page navigation.
- Assembly List page, 'Details' and tooltips.
|1.46 ||12/03/2013 || ||
|1.45 ||7/03/2013 || ||
- Remark 350 page formatting.
- Multiple assembly GET downloads.
|1.44 ||5/03/2013 || ||
|1.43 ||26/02/2013 || ||
|1.42 ||24/02/2013 || ||
- EBI and PDBe web page update.
|1.41 ||1/01/2013 || ||
- Download of multiple xml and pdb files.
- Mis-calculated dG values for uploaded files.
- Database search.
|1.34 ||3/04/2012 || ||
- CSS values mis-calculated for uploaded entries.
|1.33 ||16/03/2012 || ||
- Submission page freezes for viruses
|1.32 ||16/11/2011 || ||
- Crash (500 internal server error) when entering invalid PDB code.
|1.30 ||3/11/2011 || ||
- PISA now calculates the assembly etc on the fly for CGI GET requests, when the PDB code is not present in the database but the PDB file exists.
- Web page links made relative so that hppts requests are handled correctly.
|1.29 ||31/10/2011 || ||
- PISA query page PDB code out-of-step with internal query code. Analysis of the displayed code is automatically updated when a 4 character code is
|1.28 ||24/10/2011 || ||
- Fixed 404 error when downloading XML files using 'XML' buttons.
|1.27 ||28/09/2011 || ||
- e-mail sent to annotators concerning weekly update processing.
- PISA recompiled as a native 64-bit executable
- The correspondence between identified oligomeric
state and PQS annotation in PDB files (Remark 350)
is now reported in "Biomolecule" section of
- UniProt and SCOP ids are included as options into
PISA database searches.
- The interaction filter is introduced into PISA datbase
searches. The filter allows for the selection of PDB
entries with specific (e.g. protein-RNA) type of
- Fixed a bug that caused PISA interface searches to crash
at zero structure similarity level.
- Source code revised in order to eliminate dependence
on the computational platform, which appeared in a limited
number of instances. This does not change any previous
- Linking to PISA pages added to URL interface, see
- On request of wwPDB, the format of PDB REMARK 350
changed to include macromolecular complex
and software specifications into the output.
- Adopted changes due to the release of the
Remediated PDB Archive. PISA will continue to
accept old PDB files for processing, however PISA
database (pre-calculated results for all PDB entries,
used for fast retrieval and searches), is now built
using new, remediated, files.
- Added identification of covalent linkage in interfaces.
In assembly analysis, covalently linked interfaces are
permanently fixed except cases when it results in
- The ligand fixing algorithm is completely revised and
rewritten in order to take covalent linkage into account.
As a result, ligand fixing is done on a more rigorous
basis now, and some minor differences from previous
results (e.g. in ligand placements on assemblies) may
occur. The assembly-analysis part of PISA is now
- Surface and interface atom names are made available
for display as an option in interface tables.
- Added analysis of custom assemblies given As Is
by coordinate section in PDB or mmCIF files. This
analysis does not use crystal descriptions, therefore
it is applicable to NMR structures. The results are
found in the "assemblies" section.
- Minor interface improvements and bug fixes
- Disulphide bonds are now calculated and taken into
account at assembly analysis.
- A new facility for searching the PISA database has
been added to the service. The search options include:
- by assembly type
- by free energy of dissociation
- by assembly's accessible surface area
- by assembly's buried surface area
- by percent of buried surface area
- by presence or absence of salt bridges and disulphide bonds
- by multimeric state
- by homomeric/heteromeric assembly type
- by symmetry number
- by space group
- by presence of specific ligands
- by annotation keywords
- Fixed a bug in part simulating the crystal structure.
The bug caused crash in rare circumstances.
- Symmetry number calculations revised in algorithmic
part in order to improve results where coordinate data
are less perfect.
- Fixed a bug in part generating the visualisation data.
The bug, associated with renaming of unnamed chains,
caused crash in rare circumstances.
- Introduced a correction for interfaces that overlap
due to imperfectness of coordinate data. Total BSA is
now guaranteed not to exceed total ASA on atom level
- Fixed a bug in chemical-structural equivalencing of
NCS mates, had effect on few viral structures.
- Identification of parallel monomeric units relaxed for
non-crystallographically equivalent chains, and tightened
for NCS mates. The former is to account for imperfectness
of coordinate data, and the latter is to avoid errors in
large capsids, like 900-mer 1vb2, where monomeric units
may get almost parallel simply because of large number
- Serial numbers for symmetry operations from International
Tables of Crystallography have been dropped, as they
appear to be inconsistent through different editions of
the Tables. The symetry operations are numbered as used
by PDB at RCSB.
- Assembly output forced for all chains found in ASU.
Previously, if e.g. 9 equivalent chains A-I form
homohexamers, only one assembly A-F was shown in the
resulting tables. Now both A-F and GHIG'H'I' are shown
where G'H'I' come from neighbouring ASU.
- Control over ligand processing has been provided in
the Submission Form. This allows to exclude agents used
to aid crystallization from the processing. If not
excluded, they may lead to artifactually large assemblies
by facilitating the macromolecular binding. Cf. details
from on-line help.
- Check enforced for overlapping symmetry mates, when
they are supposed to overlap and therefore have
fractional occupancy. Previously this occasionally
resulted in multiple count of molecular interactions,
example 1CYI residue CD 302 (Cd ion).
- Fixed a job syncronisation bug causing crash before
displaying results under certain conditions
- Fixed a bug in structure search submission when one
query structure was an NCS mate, rather than original
chain given by PDB entry or coordinate file.
- Fixed a bug in secondary structure calculations,
which have caused crash on a recently deposited
- PDB/RCSB numbering for symmetry operations has been
added. The symmetry ID records now look like
S[r]_IJK, where r stands for the
PDB/RCSB serial number of the respective symmetry
- Data sent to rasmol/jmol are now uniquely named, which
overcomes cache problems caused by custom browser
and proxy settings
- Wait pages are renamed at every reload, also for
forcing browsers and proxy servers to obtain the
- Ligands are now taken into account. This changes
oligomeric states in many instances, when ligands
are found in interfaces (e.g. 1JL5).
- Fixed a bug, which sometimes caused incorrect
orthogonalisation code for uploaded structures.
- Chain IDs are now case-sensitive. Note that chain IDs
may be case-sensitive in new PDB entries.
- Identification of structurally similar assemblies now
uses a different algorithm, which makes searches for
stable assemblies considerably faster.
- Symmetry operations in output pages are now identified
also by their serial numbers in space symmetry group.
- Protein-DNA/RNA and DNA/RNA-DNA/RNA interactions
added to PISA models. PISA includes DNA/RNA-containing
- Remark 350 is written in both orthogonal and
- Generated assemblies are moved such that a maximal
number of monomeric units are found in their original
- Macromolecular interaction parameters recalibrated,
no significant changes for protein complexes.
- Minor changes of web pages include more visualisation
links and direct access to most probable assemblies
directly from the submission form.
- Parameters of chemical-thermodynamical models
have been recalibrated. This affects the estimates
of dissociation/solvation free energies, entropy
etc. No change in oligomeric states has been found
in the benchmark entries, however this does not
imply that multimeric states of all PDB entries
did not change. Change of the dissociation pattern
has been noticed in some of the benchmark entries
- A bug in the assembly enumeration procedure has been
discovered and fixed. In certain circumstances, this
bug caused a premature termination of assembly search
in a crystal, so that some of potentially interesting
assemblies with around-zero free energies of dissociation
were omitted by the search.
- The graph matching procedure used for identification
of equivalent assemblies have been optimized, which
has a drastic effect on PISA performance when
multimeric state exceeds 20.
- Structure superposition results and visualisation
of superposed structures have been added to the
"structure" section of results.
- Improved identification of chemical-geometrical
equivalences in interfaces and monomeric units.
This has had an impact on some entries where
chemically similar, but crystallographically different
chains show noticeable structural differences.
- Improved identification of casual interfaces, which
are most probably experimental artifacts. Some
entries, such as 1ir2 are affected.
- The CPU load is made more even, and a few imperfectnesses
in CPU communications cleared up, which results in
a faster completion.
- The interface search facility is enhanced by new
search options, limiting target selections to
- different similarity levels
- targets with or without multimeric assemblies
- targets with or with interfaces equivalent
to those in the query
- targets with interfaces equivalent to those
in the query, making or making no protein
contacts in (target) assemblies
- On-line documentation updated.
- Visualisation in Jmol is added.
- Buttons were added for downloading monomeric structures,
interfaces and assemblies.
- Many improvements in the identification of interface and
First public release.