PRISM A Web Server for Prediction and Visualization of Protein Protein Interactions

Tunçbağ, Nurcan
Ulubaş, Mehmet Cengız
Keskin Özkaya, Zehra Özlem
Gürsoy, Attila
PRISM is a web server for the querying, visualization and analysis of the protein interfaces and putative protein-protein interactions derived from known protein structures in PDB. Putative interactions between proteins are predicted with an efficient algorithm using structural and evolutionary similarities. The algorithm seeks possible binary interactions between proteins (targets) through similar known interfaces (templates). Template dataset is the structurally and evolutionarily representative subset of biological interfaces in PDB. Starting with all available ~50,000 interfaces as of February 2006, 8205 distinct interface clusters are generated with their representative interfaces. After elimination of the antigenantibody complexes, peptides, ligands, synthetic proteins, membrane proteins and interfaces having less than 3 hotspots – evolutionarily conserved residues on the interfaces – at each partner chain and consideration only biologically relevant interfaces, 1738 template interfaces are obtained. Target dataset is the sequentially non-redundant subset of all structures available in PDB which have less than 50 % homology. Target dataset contains 16415 structures, of which 4952 are complex structures, 11463 are monomeric structures. Surfaces of the target proteins are extracted by invoking NACCESS. If relative surface accessibility of a residue is greater than 5%, it is considered as surface residue. The prediction algorithm based on that if two proteins contain similar regions to complementary partner of a template interface, it is proposed that these two proteins interact through these similar complementary regions. After the template interfaces are split into its complementary partner chains, these partners are structurally aligned with the surfaces of the target proteins. To measure the similarity, a scoring function is used, which contains two parts; i) evolutionary similarity score and ii) structural similarity score. Evolutionary similarity includes hotspot match ratio; structural similarity part includes RMSD and residue match ratio between target protein and one partner of template interface. Prediction algorithm results in 58817 potential interactions for a score threshold of 0.85. For a biological evidence, we used the p53 (tumor suppressor protein) interaction network generated by Kohn et al., in 1999. In this network, there are 94 interactions between 55 PDB structures. We verified 84 of these 94 interactions by PRISM with a score threshold of 0.5. These verified interactions support the consistency of our prediction algorithm. Using PRISM web server, one can browse and query template, target datasets and the predicted interactions. Another feature of PRISM is the visualization of Protein Interaction Network derived from the predicted binary. This tool is based on a multi-modal network representation, incorporating the relationships between entities such as interfaces, proteins, domains, and functional annotations. This model together with its automatic layout feature results in identifying biologically important network modules.
15th Annual International Conference on Intelligent Systems for Molecular Biology (ISMB)& 6th European Conference on Computational Biology (ECCB) (21 - 25 Temmuz 2007)


HotSprint: database of computational hot spots in protein interfaces
Guney, Emre; Tunçbağ, Nurcan; Keskin, Ozlem; Gursoy, Attila (2008-01-01)
We present a new database of computational hot spots in protein interfaces: HotSprint. Hot spots are residues comprising only a small fraction of interfaces yet accounting for the majority of the binding energy. HotSprint contains data for 35 776 protein interfaces among 49 512 protein interfaces extracted from the multi-chain structures in Protein Data Bank (PDB) as of February 2006. The conserved residues in interfaces with certain buried accessible solvent area (ASA) and complex ASA thresholds are flagge...
Domain Structured Dynamics: Unpredictability, chaos, randomness, fractals, differential equations and neural networks
Akhmet, Marat (Institute of Physics Publishing (IOP), 2021-03-01)
Domain structured dynamics introduces a way for analysis of chaos in fractals, neural networks and random processes. It starts with newly invented abstract similarity sets and maps, which are in the basis of the abstract similarity dynamics. Then a labeling procedure is designed to determine the domain structured dynamics. The results follow the Pythagorean doctrine, considering finite number of indices for the labeling, with potential to become universal in future. The immediate power of the approach for f...
Out-of-Sample Generalizations for Supervised Manifold Learning for Classification
Vural, Elif (2016-03-01)
Supervised manifold learning methods for data classification map high-dimensional data samples to a lower dimensional domain in a structure-preserving way while increasing the separation between different classes. Most manifold learning methods compute the embedding only of the initially available data; however, the generalization of the embedding to novel points, i.e., the out-of-sample extension problem, becomes especially important in classification applications. In this paper, we propose a semi-supervis...
Relay Selection for Efficient HARQ-IR Protocols in Relay-Assisted Multisource Multicast Networks
Quoc-Tuan Vien, Quoc-Tuan Vien; Nguyen, Huan X.; Shah, Purav; Ever, Enver; Duc To, Duc To (2014-05-21)
This paper investigates relay selection for reliable data transmission in relay-assisted multisource multicast networks (RMMNs) where multiple source nodes distribute information to a set of destination nodes with the assistance of multiple relay nodes. Hybrid automatic repeat request with incremental redundancy (HARQ-IR) is used and supported by either a physical-layer network coding (PNC) or an analog network coding (ANC) technique employed at the relays. By deriving efficiency metrics of the HARQ-IR prot...
Multi-resolution visualization of large scale protein networks enriched with gene ontology annotations
Yaşar, Sevgi; Can, Tolga; Department of Computer Engineering (2009)
Genome scale protein-protein interactions (PPIs) are interpreted as networks or graphs with thousands of nodes from the perspective of computer science. PPI networks represent various types of possible interactions among proteins or genes of a genome. PPI data is vital in protein function prediction since functions of the cells are performed by groups of proteins interacting with each other and main complexes of the cell are made of proteins interacting with each other. Recent increase in protein interactio...
Citation Formats
N. Tunçbağ, M. C. Ulubaş, Z. Ö. Keskin Özkaya, and A. Gürsoy, “PRISM A Web Server for Prediction and Visualization of Protein Protein Interactions,” presented at the 15th Annual International Conference on Intelligent Systems for Molecular Biology (ISMB)& 6th European Conference on Computational Biology (ECCB) (21 - 25 Temmuz 2007), VIYANA, Avusturya, 2007, Accessed: 00, 2021. [Online]. Available: