Download Combinatorial docking approach for structure prediction of large

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Immunoprecipitation wikipedia , lookup

Implicit solvation wikipedia , lookup

List of types of proteins wikipedia , lookup

Rosetta@home wikipedia , lookup

Protein design wikipedia , lookup

Structural alignment wikipedia , lookup

Protein wikipedia , lookup

Bimolecular fluorescence complementation wikipedia , lookup

Protein moonlighting wikipedia , lookup

Circular dichroism wikipedia , lookup

Proteomics wikipedia , lookup

Homology modeling wikipedia , lookup

Protein folding wikipedia , lookup

Cyclol wikipedia , lookup

Protein purification wikipedia , lookup

Protein mass spectrometry wikipedia , lookup

Intrinsically disordered proteins wikipedia , lookup

Western blot wikipedia , lookup

Protein domain wikipedia , lookup

Protein structure prediction wikipedia , lookup

Nuclear magnetic resonance spectroscopy of proteins wikipedia , lookup

Protein–protein interaction wikipedia , lookup

Transcript
Greg Goldgof
March 23, 2006
CS379a Case Study
Combinatorial docking approach for structure prediction of large proteins and multimolecular assemblies
Yuval Inbar1, Hadar Benyamini2, Ruth Nussinov2,3 and Haim J Wolfson1
1 School
of Computer Science, The Raymond and Beverly Sackler Faculty of Exact Sciences, Tel Aviv University, Tel Aviv 69978,
Israel
2 Sackler Institute of Molecular Medicine, Department of Human Genetics and Molecular Medicine, Sackler School of Medicine, Tel
Aviv University, Tel Aviv 69978, Israel
3 Basic Research Program, SAIC—Frederick, Inc., Laboratory of Experimental and Computational Biology, NCI-Frederick, Bldg 469,
Rm 151 Frederick, MD 21702, USA
This paper discusses a combinatorial approach towards the docking problem for large
molecules or multi-molecular assemblies. This approach integrates existing docking programs
for pair-wise docking of smaller molecules or protein domains as well as free-energy
calculations to build determine native conformations of larger protein complexes. Such an
algorithm is much faster than trying to determine the entire protein or assembly structure through
molecular dynamics methods that compute quantum physics interactions between each atom.
Such structural information about how large proteins and protein complexes are arranged would
be of interest to biologists and could also be used in medicine to determine ways to disrupt
protein complexes in medical applications. Also, if scientists intend to perform more detailed
molecular dynamics simulations on unknown large structures having a basic framework provided
by this algorithm about where and how the proteins bind to each other may significantly decrease
the complexity of the problem. The paper also sights that knowledge of the interaction of protein
complexes could yield key information about the origin of enzymatic malfunction and disease.
I became interested in this paper because of the simplicity of the algorithm, called
CombDock (for Combinatorial Docking). For a large protein it uses a dissection algorithm to
break it down into smaller pieces. The structure of these pieces is then determined
computationally. This step is nice because many protein domains will have a known structure so
this calculation step can be skipped. The algorithm then uses an iterative combinatorial method
to combine the sections. First it screens through each possible pair of sections and determines the
likelihood of their interaction. It ranks the structure based on free energy calculations and then
takes the top N double-structures and matches them with all the remaining domains to determine
the probability of their interaction. The algorithm is greedy in that it removes the most unlikely
complexes at each iterative step. For the large molecules the program uses the protein’s basic
backbone to inform the ranking of the structures. Eventually, all the domains will be added to the
structure and the algorithm returns a list of possible structures ordered based on their probability.
Similar structures are clustered together to avoid redundancy and a final list is made.
An important note to make is that the program cannot predict whether proteins will
combine and form an assembly, but rather knowing that they do form some sort of complex can
determine its native conformation. This information is useful because biologists frequently know
that proteins interact from experiments such as immunoprecipitation, however, these experiments
provide no information on the nature of the interaction. CombDock was tested on a number of
known proteins and protein complexes and generally ranked the native conformation within the
top 10 (often 1) and usually provided results with twenty minutes. The program did not perform
as well on multi-molecular assemblies as on large molecules, presumably because knowledge of
the basic protein backbone (simply known from the sequence) provided useful information about
3D folding. The efficacy of the program was measured by the RMSD in angstroms between the
closest confirmation found and the solved structures. Since the paper was published in November
of 2005 I have was not able to find any applications of the program, but it would interest me to
see how in what ways it facilitated biological or medical research.