Email:
Phone: +972-3-640-5395
Protein Structure Analysis
MASS (Multiple Alignment by Secondary Structures)
A joint work with R. Nussinov and H. J. Wolfson
It is well established that proteins with common structural features may share similar functional properties. MASS is an algorithm for aligning multiple protein structures and for detecting common folds and spatial motifs. To date, only a few methods are available for addressing this NP-hard problem. Most of them perform a series of pairwise comparisons and thus might miss optimal global solutions. In contrast, MASS consider all the proteins simultaneously. It is a two-tier algorithm using both secondary structure and atomic representation. Exploiting the secondary structure aids in filtering out noisy results and in making the method highly efficient. The alignment is truly spatial, disregarding the sequence order of the secondary structure elements. This makes the method capable of detecting non-topological structural motifs. Another important feature of MASS is the ability to identity structural outlier proteins as well as substructures common to subsets of input proteins. Based on these capabilities, MASS can handle large-scale protein ensembles that may be heterogeneous, noisy and topologically unrelated.
EMatch - Discovery
of Atomic Resolution Structural Homologues of Protein Domains in Intermediate
Resolution Cryo-EM Maps:
A joint work with K. Lasker, M. Shatsky,
R. Nussinov and H. J. wolfson
Cryo-EM is a powerful technique for elucidating structural information on large macromolecule assemblies that cannot be determined at atomic resolution. We develop EMatch with the desire to bridge the resolution gap by combining cryo-EM maps of protein assemblies with atomic resolution data on single domains. The input is an intermediate-resolution cryo-EM map of a protein assembly. The rationale is that at this resolution secondary structure elements become apparent. EMatch detects helices in the given map and uses their spatial arrangement to query a dataset of known atomic domains to find the ones that can be fit into the map. The result is a quasi-atomic structural model of the assembly. EMatch has been successfully tested on simulated data as well as on experimental cryo-EM data of native GroEL at 6 Å resolution.
Structural
Similarity in Genetic Interactions
A joint work with A. Shulman-Peleg, D.
Schneidman-Duhovny, R. Nussinov, H. J. Wolfson and R. Sharan
A genetic interaction (GI) between two genes is defined when mutations of both genes have a combined effect not exhibited by either mutation alone. The recent availability of large-scale GI networks in yeast and worm allows the investigation of the biological mechanisms underlying these interactions at a global scale. To date, less than 2% of the known GIs in yeast or worm can be accounted for by sequence similarity. We performed a genome-scale structural comparison among protein pairs in the two species. We found that significant fractions of GIs involve structurally similar proteins, spanning 7% and 14% of all known interactions in yeast and worm, respectively. By analyzing this data, we observed that structure similarity information is more predictive of GIs than sequence information. Moreover, a combination of structural and functional information is more indicative of GIs than either property by itself. Based on these observations, we suggest a putative mechanism for genetic interactions among structurally similar proteins.
ARTS - Alignment of RNA
Tertiary Structures
A joint work with R. Nussinov and H. J.
Wolfson
Much like proteins, structural similarities shared by RNA molecules may imply a similar function. Nevertheless, the current existing tools for RNA structural comparison provide only partial solutions. These tools either work at the secondary structure level or are suitable for detecting predefined or local contiguous tertiary (3D) motifs only. In contrast, ARTS is the first method that performs a structural comparison that is truly three-dimensional and irrespective of the order of the nucleotides along the chain. ARTS is thus suitable for detecting a-priori unknown common 3D substructures that may not necessarily be contiguous. These substructures can be large global folds with hundreds and even thousands of nucleotides as well as small local motifs. The method is also highly efficient and suitable for searching large databases.
DARTS - Alignment of RNA
Tertiary Structures
A joint work with M. Abraham, R. Nussinov
and H. J. Wolfson
With the aim of addressing questions relating to RNA structural diversity, we have examined the conservation of RNA at three structural levels: sequence, secondary structure and tertiary structure. In addition, we have developed a fully automated method for clustering known RNA structures mainly based on their spatial resemblance. Applying the method to all known RNA structures resulted in a classification database named DARTS. The classification is hierarchical reflecting the similarity relationship between the structures. The 94 different clusters reveal the current fold repertoire of RNA, where structural similarities within and between clusters may suggest possible local tertiary motifs. In addition, DARTS is provided with a search tool for comparing a newly determined RNA structure with the ones in the database. This allows RNA experts to gain more insights into their structures.
PharmaGist - Pharmacophore
Detection via Multiple Flexible Alignment of Drug-like and
Applications for Virtual Screening
A joint work with D. Schneidman-Duhovny, Y. Inbar, R. Nussinov and H. J. Wolfson
Pharmacophore is the spatial arrangement of features that is essential for a drug-like molecule (ligand) to interact with a specific target receptor. A major challenge in rational drug design is to detect pharmacophores in cases where the 3D structure of the receptor is unknown and only a set of known active ligands is available. PharmaGist is a novel method for tackling this task by performing multiple flexible alignments of the input ligands. The main innovation of this approach is that the flexibility of the ligands is handled explicitly and in deterministic manner within the alignment process. Other key advantages of PharmaGist are the high computational efficiency and the ability to detect pharmacophores common to subsets of input ligands, a characteristic that makes the method tolerant to outliers and to several binding modes. An important application of PharmaGist is for virtual screening, namely the identification of potential new drugs from a large database of chemical compounds based on their alignment with a pharmacophore model.