Protein Structure Bioinformatics Resources

Protein Structure Bioinformatics Resources
An arbitrary subset gathered by Eric Martz
provided as a supplement to Protein Explorer.
EMail suggestions to me.

[PE] marks resources linked into Protein Explorer in context-sensitive help for certain operations, or in the External Resources window (accessed from the PE Site Map).

123D+, threads a sequence through a set of 3D structures. It combines sequence profiles, secondary structure prediction, and contact capacity potentials to find the most compatible fold among the 3D structures, and the best alignment of the sequence with that fold.
3D-Fun (available from Netherlands or Poland), accepts protein coordinates in the standard PDB format and compares them with all known protein structures by 3D structural superposition. Structural hits having known functions are listed. Thus, 3D-Fun attempts to predict function from structure.
3D-PSSM, attempts to predict 3D structure and probable function from a protein sequence in part by threading with known structures and scoring for compatibility.
BioInfoBank Meta Server assists in the construction of comparative (homology) models.
Biology Workbench [PE]: Integrated access to sequence and structure databases with result management tools. Very powerful but not very user friendly. Instructions for use provided in conjunction with Protein Explorer's MSA3D capability.
Books on protein structure and crystallography.
Columba-DB: Protein Structure Annotation. This meta-server provides an integrated summary with links to details for information from the PDB, KEGG, ENZYME, ExPASy, DSSP, CATH, SCOP, SwissProt, NCBI Taxonomy, GO, and PISCES.
CAPRI: Critical Assessment of PRedicted Interactions, to assess the capacity of protein-protein docking methods to predict protein-protein interactions.
CASP: Critical Assessment of techniques for protein Structure Prediction. An annual competition in which information is solicited from X-ray crystallographers and NMR spectroscopists on structures about to be solved. Before the empirical structures are known, modelers submit structure predictions in three categories: comparative modeling, fold recognition or threading, and ab initio folding. Includes Ten Most Wanted, to provide structural and functional insights into biologically important proteins, particularly those that are intractable to experimental structural determination.
CASTp: A server that identifies pockets and cavities in proteins, and quantitates their volumes. Atoms lining each pocket or cavity can be displayed in Chime, RasMol, or MAGE.
Cavities and pockets in proteins, see CASTp.
Combinatorial Extension, Finding 3D similarities in protein structures without reference to sequence. Automated intelligent alignment of protein chains (ligands are discarded).
Comparative Modeling (Homology Modeling) servers are continuously and automatically evaluated by EVA. There is also a structure prediction meta-server for difficult cases, the BioInfoBank Meta Server. For straightforward cases, comparative modeling is automated by SWISS-MODEL.
Conformational Epitope Prediction Server, Predicts possible antigenic epitopes on surfaces of protein antigen structures submitted. Displays predicted epitopes in Jmol.
ConSurf [PE]: Automated identification of functional regions in proteins of known 3D structure based upon evolutionary conservation.
Dali: Compares structure of submitted coordinates with those in the Protein Data Bank, without reference to sequence. Returns alignment of structure neighbors. For comparisons within the PDB, use FSSP. See Structure Searching.
Database of Macromolecular Movements: Categorizes types of movements in conformational changes of macromolecules. Offers a morph server to animate conformational changes, and an archive of morphs generated by visitors, with numerous viewing options. Such morphs can be animated in Protein Explorer.
DeepView: a free, powerful molecular modeling program.
DisEMBL: an Intrinsic Protein Disorder Prediction server. "DisEMBL is a computational tool for prediction of disordered/unstructured regions within a protein sequence. ... Avoiding potentially disordered segments in protein expression constructs can increase expression, foldability and stability of the expressed protein." See also disprot.org.
DisProt.org: "Database of Protein Disorder (DisProt) is a curated database that provides information about proteins that lack fixed 3D structure in their putatively native states, either in their entirety or in part." See also DisEMBL.
Entrez, a powerful system for integrated searching of literature, sequences and genomes, structure, and other databases. See NCBI.
EVA continously and automatically analyses protein structure prediction servers in 'real time'. Evaluates servers for secondary structure prediction, comparative modeling, inter-residue distances and contacts, and threading.
Function prediction servers for protein include:
- 3D-PSSM (from sequence).
- 3D-Fun (from structure).
Functional Site prediction servers for protein include:
- ConSurf.
- Hotpatch.
FSSP: Fold classification based on Structure-Structure alignment of Proteins: Returns structural alignments and structurally-based sequence alignments for structure neighbors in the PDB. For structures not in the PDB, see Dali.
Glossaries include
- Protein Explorer's Help/Index/Glossary (Martz, University of Massachusetts)
- Glossary of Terms from Crystallography, NMR, and Homology Modeling (Rhodes, University of Maine)
- For other bioinformatics glossaries, search Google
GRASS, the Graphical Representation and Analysis of Structure Server, using the widely acclaimed GRASP software, renders a PDB file in a manner selected by the user from menus, and offers visualization in Chime or VRML as well as Grasp. Notably, molecular surfaces can be colored by electrostatic potential, hydrophobicity, distances from a ligand, surface curvature
Helmholtz Network for Bioinformatics: an integrative web portal for bioinformatics resources, including protein structure prediction, alignment generation, threading, and structure-similarity searching (all under Protein Structure).
HighWire, full-text searching of journal articles.
HIV Macromolecular Structures, a component of the Sequence Database of the HIV Databases at Los Alamos National Laboratory, Los Alamos, New Mexico USA.
Homology Modeling: see Comparative Modeling.
"HotPatch finds unusual patches on the surface of proteins, and computes just how unusual they are (patch rareness), and how likely each patch is to be of functional importance (functional confidence (FC)). The statistical analysis is done by comparing your protein's surface against the surfaces and functional sites of a large set of proteins, with known functional sites."
Journals covering macromolecular structure.
Journal articles on major principles of protein structure.
KFC predicts "hot spots" in protein-protein interfaces, namely, the subsets of contacting residues responsible for most of the interfacial binding free energy. KFC uses shape specificity, atomic contacts, and noncovalent bond analysis to to make its predictions. Site-directed mutation of hot spots is an effective means of disrupting a protein-protein interaction. KFC is an alternative to costly experimental identification of hot spots.
Map2Model is a server for evaluation of crystallographic models and their agreement with electron density maps. It can remove atoms of side chains that are not consistent with the maps, remove improperly placed water molecules, check the B-factors of atoms in the provided model, and compare R and R-free values obtained as the result of refinement with the averages characteristic for the data resolution shell.
MaxSprout generates protein backbone and side chain co-ordinates from a C(alpha) trace.
Membrane Protein Data Bank has many membrane-protein specific search capabilities.
Membrane Protein Structure Databases:
- Membrane Protein Data Bank
- Orientations of Proteins in Membranes
- Transmembrane Protein PDB
MolProbity [PE] conducts a new and powerful structure validation based on all-atom contacts analysis, producing automatic corrections (rotamer flips). The flips can be inspected and accepted or rejected in a java-based Kinemage-style viewer. The corrected PDB file, containing added hydrogen atoms, is provided. For secure use, free software that performs the all-atom contacts analysis can be downloaded and run locally. The server also offers very useful Ramachandran plots (identifying residues outside the favorable regions), C-beta deviations, and rotamer analysis. See All-Atom Contacts and Quality of the molecular model.
National Center for Biotechnology Information (NCBI), includes Entrez, PubMed, and a Structure database (see Structure Searching).
NMR, Most Representative Model: This server identifies the most representative model in the ensemble of models resulting from NMR experiments. The most representative model is identified by Olderado (which appeared to be out of service when this entry was added).
OCA [PE], a powerful alternative mechanism for searching the world structure database in the Protein Data Bank.
Olderado: see NMR.
Orientations of Proteins in Membranes (OPM) "provides hydrophobic thicknesses and orientations of membrane proteins with respect to the hydrocarbon core of the lipid bilayer." Offers visualization of PDB files with the predicted membrane surface planes indicated by heteroatoms.
PDBLite [PE] offers a simple interface for searching the World Wide Protein Data Bank. PDB Lite is designed for students, educators, and non-specialists. Clear help is provided, crystallographic details are omitted, and jargon is explained.
PDB_REDO re-refines models from X-ray diffraction data, sometimes improving the model when compared to the original published model.
PDB_Select: generates sequence nonredundant subsets of chains in the Protein Data Bank using methods of Hobohm, Sander et al.. Used by OCA for nonredundant subsets.
PDBSProtEC: A resource to link PDB chains with SwissProt codes and EC numbers. Enter one of a PDB identification code, an EC (Enzyme Commission) number, or a Swiss-Prot identification code(s), and get the other two. Updated daily.
PDBSum: a pictorial database providing an at-a-glance overview of every macromolecular structure deposited in the Protein Data Bank. Includes Jmol view and rotatable (RasMol or Chime) images of surfaces and clefts in surfaces, and a Raster3D image generating form. Highlights lists the oldest/newest PDB entries, and those with the largest/smallest molecules, most chains, longest chains, most/largest ligands, highest and lowest resolution, etc.
PISCES: Guoli Wang and Roland Dunbrack's site that culls entries from the PDB based on resolution and sequence identity, among other things. See also PDB_Select.
PMut: predicts liklihood of disease resulting from mutations of amino acids in proteins. To display disease-probability as colors in Chime, go to Precalculated PMut, submit a PDB code (for example, 1IRD for human hemoglobin), and click on the Structure links.
Pockets and cavities in proteins, see CASTp.
Probable Quaternary Structure (PQS) [PE] uses an empirically derived weighted score to estimate whether protein-protein contacts that occur in crystals are specific oligomer interfaces or crystal artifacts. For those deemed specific oligomers, the oligomer is provided. When multiple copies of the molecule occur in the asymmetric unit, single copies are provided. PQS also provides complete virus capsids, using symmetry operations typically provided by the authors in the PDB file (virus capsid examples). See PE's Introduction to Probable Quaternary Structure.
PROCAT, a database where you can search for 3D enzyme active site template motifs in a protein structure.
ProMode calculates vibrational dynamics of protein molecules and among other things, displays the results as an animation in MDL Chime.
ProSAT: maps ProSite patterns and SWISS-PROT features onto 3D protein structures.
Protein Data Bank [PE], the sole primary international repository of all published macromolecular 3D structures. Includes extensive links to related bioinformatics resources.
PubMed [PE], searchable journal article abstracts from the US National Library of Medicine. See also HighWire and NCBI.
RDP, threading by recursive dynamic programming, available through the Helmholtz Network (under Protein Structure).
S2C:
- At OCA: (updated weekly, fast response; find PDB file, then under Data Retrieval, click on SEQRES to Coordinates)
- Fox Chase Cancer Center (original development site by Wang and Dunbrack; updated irregularly; slow response).
"A database correlating sequence and atomic coordinate residue numbering in the Protein Data Bank". Shows alignment of SEQRES vs. sequence with coordinates in ATOM records. Easiest way to spot gaps (amino acids in the crystallized compound that lack coordinates), but doesn't do nucleotides, and does not show sequence microheterogeneity (e.g. 1CBN).
SPAM: Systematic Protein Annotation and Modeling is a multi-institutional initiative to provide tools to assist in selecting and registering targets for structural genomics and for associating structure with function.
STAN: STructure ANalysis server at University of Uppsala, Sweden. Includes Ramachandran analysis and routines to detect water molecules that should be metal cations, and trans peptides that should be cis.
Structure searching & comparision: finding structure neighbors without reference to sequence: See
- DALI (for unpublished structures)
- Combinatorial Extension.
- FSSP (neighbors within the PDB)
- VAST (Vector Alignment Search Tool) from NCBI.
Structural Classification of Proteins (SCOP). See very brief overview by Chothia et al..
SuMo: Surfing the Molecules. Automated search for similar sites in proteins, "to screen the Protein Data Bank (PDB) for finding ligand binding sites matching your protein structure or inversely, for finding protein structures matching a given site in your protein. This method is neither based on amino acid sequence nor on fold comparisons. Priority is given to biological relevance." SuMo result for 1LPM (entire protein).
SWISS-MODEL [PE], "an automated comparative protein modelling server". See Introduction to Comparative Modeling.
Swiss-Prot, "a curated protein sequence database which strives to provide a high level of annotations (such as the description of the function of a protein, its domains structure, post-translational modifications, variants, etc.), a minimal level of redundancy and high level of integration with other databases." To get Swiss-Prot codes from a PDB code (or EC number), use PDBSProtEC.
Threading servers include 123D and RDP available through the Helmholtz Network (under Protein Structure), and 3D-PSSM. Threading servers are continuously and automatically evaluated by EVA.
TOPS Services At Glasgow University: Protein topology diagrams, topology matching between domains, structure comparisons, motif matching.
Transmembrane Protein PDB. Identifies the transmembrane proteins in the PDB. Searchable, browsable, downloadable.
Visualization Servers: Listed at molvisindex.org. These servers assist in generating visualizations of any model (PDB file) you submit. They require much less technical knowledge than does direct use of visualization software packages.
Visualization Software: See molvisindex.org.
WhatIF WWW Interface by Gert Vriend offers some powerful structure validation and modeling routines, including addition of hydrogens, and crystal contacts.
World Index of Molecular Visualization Resources: molvisindex.org [PE]. Visitors contribute links to on-line resources in categories including ready to use tutorials on specific molecules (including proteins, organic chemicals, inorganic crystals), K12 resources, resources in German, French, Italian, Japanese, Spanish, etc., technical resources on how to create Chime websites, DeepView (SwissPDB-Viewer) resources, sources of molecules (PDB files), galleries, email lists, free and commercial software. Over 200 entries are described in detail and cross-indexed by subjects and authors.