Protein Research Tools
links: Protein structure, function and analysis
[Home] [Search Engines] [RNA info]
2009.02.27: I've made some attempt to up date this page. Some information should be possible to find again.
Protein stucture/function resources
data base resources [top]
databank for protein structures found via x-ray crystallography
databank for protein structures found via NMR spectroscopy
Southwest Biotechnology and Information center
SWBIC - Bioinformatics (no longer so helpful)
This used to be a real database of information and very well curated, but now
it maybe has some applications. However, when I tried to used them, they didn't
work, so maybe it doesn't even have that.
Basically a page like this one, but probably better maintained.
Some useful links (also a mirror site for PredictProtein).
mutation database (1) (he's dead jim)
Gives the thermodynamic parameters for mutations of a large number of
proteins. Probably quite useful for protein engineering and protein
structure prediction.
Homology and Threading Prediction [top]
ExPASy (Switzerland mirror) (homology modeling)
A package of programs linked together to help search
for secondary structure, homology, and 3D structure.
This server does not mess with structural homologies less than 25%,
so you will not get any answer for your twilight zone structures.
Also provides numerous molecular biology resource links.
Other mirror sites: [Australia] [Canada] [China] [Korea] [Taiwan]
Phyre (threading) and 3D-pssm (an older version of the same concept).
This is a protein threading program which does help when your homology is less
than 25%. If you find yourself in the twilite zone where standard threading
(e.g., Swiss-model) cannot help you, my own recommendations are as follows:
first try to get some idea of the general structure of the protein you are
trying o solve, submit queries of moderate size (maximum 100 aa), finally,
look thoughtfully at the results. The best fits will probably be in with the
better scores, but that does not mean that it will be the "best" score.
Psi Predict (threading)
According to the authors "The PSIPRED protein structure prediction server incorporates
two new methods (PSIPRED and GenTHREADER) and one established method (MEMSAT 2) for
predicting structural information about any given protein from its amino acid sequence alone."
SAM-T08 (remote homology detection server)
Uses a hidden Markov model do sequence/structure alignments
of a protein. Requires FASTA formated files for submission.
Provides an nice visual alignment of the residue substitution.
The format used in the homology modeler is harder to evaluate
compared to 3Dpssm or PSIpredict. Predictions are comparable
to PSIpredict. This program does homology modeling even in
the twilight zone. Also can use SAM-T02 (remote homology) or SAM-T99.
123D+ (threading)
This server does not seem to be actively supported, but it is a
threader and it's still operational.
Secondary structure prediction [top]
Jpred is a web server that takes a protein sequence or multiple
alignment of protein sequences, and from these predicts secondary
structure using a neural network called Jnet. The prediction is the
definition of each residue into either alpha helix, beta sheet or
random coil secondary structures. These results are aligned with
predictions from pssm and hmm models.
In some ways, I was rather impressed with this model at least for
the problems that have interested me so far. As with most secondary
structure servers, you have to sift, but my feeling is that it does
a little bit better than PSIpredict on the secondary structure
prediction. However, you should consider for yourself using McGuffin's
(aka PSIpredict) continuously updated prediction success [evaluator].
An all-around powerful package.
The protein prediction server (EMBL).
PredictProtein is a neural network secondary structure prediction and sequence
analysis package. This is a full integrated package and works fairly well.
See also validation for NMR constrant data: [Validation Suite].
The program seems to work ok. In my own problems, it actually
predicted one segment of a secondary structure that other programs
missed. It also missed a segment that some others got, but anyway.
Secondary structure predition server. This is a rather early
version of neural network based secondary structure prediction
and it is gradually becoming obsolete. Still, it is good to
compare results (even Chou-Fasman) because it gives some idea
about the differences between these models.
GOR V(still rather funky)
cdm(works)
GOR IV(this one works)
-----------------------------------------------------------------------
2004.03.22: Found a website where I can access GOR IV ("this one works").
The following are somewhat related programs written by the same [group]:
[QL with homologues] and [QL without homologues] .
-----------------------------------------------------------------------
Neural networks are perhaps a little bit more powerful (probably 5-10%).
However, the strong points of the GOR model are as follow:
(1) It relies on statistical parameters to make predictions. In this
respect it captures part of the essense of a real statistical mechanical
model, although the parameterization is strictly obtained from the
observed frequency of certain types residues that correlate well with
a given secondary structure.
(2) It is not a black box like a neural network. In this regard, as we
improve our understanding of how to identify structure, the GOR model
has much better pedigogical appeal because it can teach us why
a given approach or parameterization works or not.
The title means the advanced form of "self-optimized prediction method" (SOPM).
I presently have few details on this method. It appears to be a little bit
along the same lines of the GOR method.
NNSSP server(probably not functional)
I don't know much about this one. It is based on nearest neighbor
secondary structure prediction, but I was not able to get it to
work myself. It is recommended by the GOR IV group, as a prediction
software server, and maybe something didn't get entered correctly. Anyway...
Sequence alignment servers and other functions [top]
This is a standard place to start the search. Multiple sequence
alignments are done by PSIblast.
ClustalW does multiple sequence alignments and constructs a
tree of the evolution of the sequences. Important: ClustalW
can read many kinds of file formats, however, the input file
require a particular suffix for the program to successfully read
the file.
According to the authors: "MPSA is a stand-alone software package
for protein sequence analysis with a high integration level and Web
client/server capabilities. It provides many methods and tools, which
are integrated into an interactive graphical user interface."
SAM-T08 server
Sam-T08 does more than alignments. It is basically a kind of
homology prediction program. It uses a hidden markov model
to obtain the alignments and structures.
Easy ToPLign Fold Prediction(links currently broken)
According to the authors: ToPLign implements standard pairwise and multiple
alignment methods with flexible parameter handling.
FlexX and FlexS (commercial SW)
FlexX is a computer program for predicting protein-ligand interactions.
FlexS is a computer program for predicting ligand superpositions.
Protein sequence multiple alignment tools using statistical rules.
LOCK (link seems to be broken)
Hierarchical Protein Structure Superposition. Tries to match
the input structure with other structures in the PDB. It seems
to be part of an old program called foldminer
that is still available.
sequence analysis (Baylor College of Medicine)
A little hard to catch onto at first, but a general sequence analysis
program for studying protein and RNA structure and function. Seems
to be not so well maintained and basically combines the results of
various other servers where you can request this already.
Predicts the presence and location of signal peptide cleavage
sites in amino acid sequences from different organisms: Gram-positive
prokaryotes, Gram-negative prokaryotes, and eukaryotes. The method
incorporates a prediction of cleavage sites and a signal peptide/non-signal
peptide prediction based on a combination of several artificial neural networks.
ANTHEPROT (ANalyse THE PROTeins). The main idea behind this package is
"to integrate into a single package most of the methods designed for protein sequence analysis."
Package no longer on a server.
Structure Classification Tools (for Proteins) [top]
No server available, but it provides you with information
on how to classify the loops of proteins.
secondary structure (stride server) (He's dead Jim).
server. Information on the software can be found at [Stride].
secondary structure (DSSP) (seems he's dead Jim.)
No server available but you can download the program and use it
after signing a license agreement. This seems to be considered the
standard for classification, but stride is provided on a server and
should predict similar classification.
Detects membrane segments, signal peptides, anchors and coiled coils.
Fold Recognition Servers [top]
Phyre fold recognition server (threading) and 3D-pssm (an older version of the same concept).
This site can make a stab in the dark for you if you are pushing
the trottle button to the twilight zone (homology less than 25%).
The results cannot be guaranteed, and certainly require considerable expert
evaluation that is far from "automatic", but you can find some leads at least.
Psi Predict (Fold recognition server)
Like 3D-pssm, this approach can also make a stab in the dark for you
if you are pushing your luck to the twilight zone. Again, the
results cannot be guaranteed, and certainly require considerable expert
evaluation that is far from "automatic", but you can find some leads at least.
bioinbgu Main Page (fold recognition server)
Ben Gurion University (Dept of Math and CS) protein fold recognition server.
NOTE: to past in a sequence, you must use the following format:
..
my sequence
--
and the sequence must be longer than 40 characters: else it does not work.
090227wd: The site has probably been disabled long ago. There is another program
available Meshi that may be useful.
UCLA-DOE FOLD Server (fold recognition server)
YAFRS: yet another fold recognition server.
There are also a variety of other services including structure validation,
alignment and some related software applications. [DOE software at UCLA].
SWISS-MODEL (homology modeler)
The requirements at this site are very stringent for homology modeling.
The protein should have at least 25% homology with existing and solved
protein structures. Otherwise, well, you're out of luck in this bar pal.
Not really intended for "fold recognition" per se, but if you have a good
structure, this will recognize the fold.
Dali (PDB file analyzer)
Distance-matrix ALIgnment: you can submit a structure and DALI
and DALI will supply a set of similar structures and their alignments
with that query structure.
123D+ (threading)
This server doesn't seem to be actively supported, but it's still operational.
123D+ is a program which combines sequence profiles, secondary structure
prediction, and contact capacity potentials to thread a protein sequence
through the set of structures.
SARF stands for Spatial ARangement of backbone Fragments.
Protein Imaging Software [top]
Linux software for chemistry: molecular modeling, visualization,
graphic, quantum mechanic, dynamic, kinetic, simulation
This is a nice viewer for PDB formated files.
It also has some other features such as reasonably
simple ways to do mutations of the structure. [tutorial] introduction.
Mage, another pdb viewer [home page]
Jane Richardson has a lot of artistic figures of protein structure
and taxonomy in the literature. The structure of this program is a
little more complicated then some of the others.
Deep View is much easier to use, and probably more versitile, but
rasmol is a historical standard.
other molecular visualization software
The list is getting a bit old and some of the links do not work.
alignment software.
DOE servers (and software) at UCLA
Actually a number of things are located here including related software, a
Fold recognition, Verify3D Evaluation, ERRAT evaluation, and Crystal Twinning.
links to protein structure/analysis SW
Maintained by PDB (protein data bank). Fairly well curated and
contains a lot links to useful information that I have not found
by the usual trial-and-error search-engine approach.
Modules written in perl for doing bioinformatics etc. [documentation].
Transmembrane helix prediction [top]
TransMembrane (hidden markov model) TMHMM
prediction standards [top]
If you'd like to figure out a structure, here's a place that is
willing to give you one to solve.
What_if (What_if, prove, procheck)
This is a fairly broad and general resource that looks at a lot of
issues related to the quality of the determined protein structure.
The combined results of what_if, prove, and procheck help to identify
places in the structure prediction that may have some problems.
Structure Validation Tools (Richardson lab)
Also a useful site for checking a predicted protein structure.
Provides warnings for particularly out of rule residues in a
graphical format. Jane Richardson is famous for the remarkable
figures found in the "protein structure taxonomy" in Advances in
Protein Chemistry series.
Validation tools (DOE-UCLA project)
YAPSVT: Yet Another Protein Structure Validation Tool
Prostar (NIST)
Promises to be a site for setting standards for protein
potentials used to evaluate and predict protein structures.
Also provides a decoy set of structures for testing. It is
not so well maintained lately, probably a consequence of
the difficulty of the problem and the voluntariness of the
submission process.
Valuable web resource information [top]
circular dichroism (CD) spectroscopy
Brief but informative essays on the technique, strengths and limitations
of CD spectroscopy in relation to protein structure determination.
original address (broken link)
Put on by the folks at Swiss-Prot. Actually very informative for
beginners with a lot of straight forward comments about the strengths
and limitations of protein structure prediction and determination.
The graphics are quite remarkably good and very readable.
provides all sorts of links to various information.
I'm sorry, as a fellow musician, I just could not resist.
For me, I prefer to work it out the old way, at the keyboard
in some improvisation, but I respect that there are other
ways to create music, and this is a somewhat popular pathway
for 20th century music.
last update: 110128