Bujnicki lab - Statistical potentials
  • SOFTWARE
    • DATABASES
      • REPAIRtoire

        • REPAIRtoire is an online database for systems biology of DNA damage and repair. The database collects and organizes the following types of information: (i) DNA damage linked to environmental mutagenic and cytotoxic agents, (ii) pathways comprising individual processes and enzymatic reactions involved in the removal of damage, (iii) proteins participating in DNA repair and (iv) diseases correlated with mutations in genes encoding DNA repair proteins. The DNA repair and tolerance pathways are represented as graphs and in tabular form with descriptions of each repair step and corresponding proteins, and individual entries are cross-referenced to supporting literature and primary databases. In addition, a tool for drawing custom DNA-protein complexes is available online.

      • MODOMICS

        • MODOMICS is the first comprehesive database system for systems biology of RNA modification. It integrates information about the chemical structure of modified nucleosides, their localization in RNA sequences, pathways of their biosynthesis and enzymes that carry out the respective reactions (including the available protein sequence and structure data). It also provides literature information, and links to other databases. The database can be queried by the type of nucleoside (e.g. A, G, C, U, I, m1A, nm5s2U etc.), type of RNA, position of a particular nucleoside, type of reaction (e.g. methylation, thiolation, deamination, etc.), and name or sequence of an enzyme of interest. Options for data presentation include graphs of pathways involving the query nucleoside, multiple sequence alignments of RNA sequences and tabular forms with enzyme and literatiure data. The contents of MODOMICS can be accessed through the World Wide Web at http://genesilico.pl/modomics/

      • RNA Bricks

        • RNA Bricks is a database of RNA 3D structure motifs and their contacts, both with themselves and with proteins. The database provides structure-quality score annotations and tools for the RNA 3D structure search and comparison. 

      • RNArchitecture

        • RNArchitecture is a database that provides a comprehensive description of relationships between known families of structured non-coding RNAs, with a focus on structural similarities. The classification is hierarchical and similar to the system used in the SCOP and CATH databases of protein structures. Its central level is Family, which builds on the Rfam catalog, and gathers closely related RNAs. Consensus structures of Families are described with a reduced secondary structure representation. Evolutionarily related Families are grouped into Superfamilies. Similar structures are further grouped into Architectures. The highest level, Class, organizes Families into very broad structural categories, such as simple or complex structured RNAs. Some groups at different levels of the hierarchy are currently labeled as “unclassified”. For each Family with an experimentally determined 3D structure(s), a representative one is provided. RNArchitecture also presents theoretical models of RNA 3D structure and is open for submission of structural models by users. Compared to other databases, RNArchitecture is unique in its focus on structure-based RNA classification, and in providing a platform for storing RNA 3D structure predictions. RNArchitecture can be accessed at http://iimcb.genesilico.pl/RNArchitecture/.

          *The classification is expected to evolve as new data become available and we would like to encourage all interested parties to submit their suggestions for improvement as well as 3D structure predictions* 

    • STAND-ALONE
      • PROTMAP2D

        • Two-dimensional maps of contacts summarize interactions between amino acids in the structure. They reveal characteristic patterns of interactions between secondary and super-secondary structures and are very attractive for visual analysis. The overlap of the residue contact maps of two structures can be easily calculated, providing a sensitive measure of protein structure similarity. 

      • ModeRNA

        • We developed a method for 3D homology modeling of RNA structures. It requires a pairwise sequence alignment and a structural template to generate a 3D structural model of the target RNA sequence via either a fully automated or script-based approaches. ModeRNA is capable of handling 115 different nucleotide modifications and bridging gaps using fragments derived from an extensive fragment library.

      • RNAmap2D

        • RNAmap2D is a software tool for calculation of contact and distance maps based on user-defined criteria, and to some extent, quantitative comparison of pairs or series of contact maps and visualization of the results.

      • FILTREST3D

        • Filtrest3D is a program for discrimination of a large number of alternative models of protein structure or protein-ligand structure against a set of restraints derived from low-resolution experimental analyses (such as cross-linking, mutagenesis, circular dichrosm etc.) as well as from computational predictions (e.g. solvent accessibility, amino acid contact maps).

      • PyRy3D

        • PyRy3D is a software tool for modeling of structures for large macromolecular complexes. It uses Monte Carlo simulation to sample conformational space and to identify the best fit of complex components structures into a density map. Complex building process is based on distance restraints derived from experiments. 

      • Statistical potentialsRNA-Protein docking

        • We developed two medium-resolution, knowledge-based potentials for scoring protein-RNA models obtained by docking: the quasi-chemical potential (QUASI-RNP) and the Decoys As the Reference State potential (DARS-RNP). Both potentials use a coarse-grained representation for both RNA and protein molecules and are capable of dealing with RNA structures with posttranscriptionally modified residues. In our tests that compared these methods to other published potentials, DARS-RNP showed the highest ability to identify native-like structures.

      • SimRNA

        • SimRNA - a tool for simulations of RNA conformational dynamics, including RNA 3D structure prediction

      • SupeRNAlign

        • SupeRNAlign is a tool for flexible superposition of RNA 3D structures. Our program implements an iterative algorithm that splits RNA structures into fragments and superimposes them using existing tools (currently R3D Align). Finally, the superimposed structures are saved to a .pdb file and a sequence alignment is generated.

      • QRNAS

        • QRNAS is an extension of the AMBER simulation method with additional terms associated with explicit hydrogen bonds, co-planarity base pairs, backbone regularization, and custom restraints. QRNAS is capable of handling RNA, DNA, chimeras and hybrids thereof, and enables modeling of nucleic acids containing modified residues. 

    • SERVERS
      • SimRNAweb

        • SimRNAweb is a web server interface to the SimRNA method for RNA 3D structure modeling method

      • MetaServer

        • The GeneSilico MetaServer is a gateway to a number of third-party methods for protein structure prediction (identification of domains, secondary structure prediction, fold-recognition, and finally, 3D model generation). Users can submit a protein sequence (or alignment) by a single click, then analyze the summary of results generated by many methods, and finally predict the protein structure according to the "consensus" approach.

      • FILTREST3D

        • This is an "alpha" version of a server for discrimination of a large number of alternative models of protein structure against a set of restraints derived from low-resolution experimental analyses (such as cross-linking, mutagenesis, circular dichrosm etc.) as well as from computational predictions (e.g. solvent accessibility, amino acid contact maps).

      • CompaRNA

        • The CompaRNA web server provides a continuous benchmark of automated standalone and web server methods for RNA secondary structure prediction. It has been inspired by the EVA and Livebench servers for benchmarking of protein structure prediction tools, which have greatly contributed to progress in the field of structural bioinformatics. The aim of CompaRNA is to assess the state of the art in RNA structure prediction, provide a detailed picture of what is possible with the available tools, where progress is being made and what major problems remain. The CompaRNA server is a valuable resource for all researchers who focus their attention on the usage and development of RNA structure prediction methods.

      • ModeRNA server

        • ModeRNA server is an online tool for RNA 3D structure modeling by the comparative approach, based on a template RNA structure and a user-defined target-template sequence alignment. It offers an option to search for potential templates, given the target sequence. The server also provides tools for analyzing, editing and formatting of RNA structure files. It facilitates the use of the ModeRNA software and offers new options in comparison to the standalone program.

      • MetalionRNA

        • MetalionRNA is a web server for the prediction of metal ions (magnesium, sodium, and potassium) in RNA 3D structures, based on a statistical potential inferred from the analysis of binding sites observed in experimentally solved RNA structures. The server is also capable of predicting Mg2+-binding sites for DNA structures.

      • MetaLocGramN

        • The MetaGramLocN is a method for subcellular localization prediction of Gram-negative proteins. The MetaGramLocN is a gateway to a number of primary prediction methods (various types: signal peptide, beta-barrel, transmembrane helices and subcellular localization predictors). The MetaGramLocN integrates the primary methods and based on their outputs provides overall consenus prediction. To make a prediction for your protein sequence use Submit or SOAP client In our benchmark, the MetaLocGramN performed better in comparison to other SCL predictive methods, since the average Matthews correlation coefficient reached 0.806 that enhanced the predictive capability by 12% (compared to PSORTb3).

      • RIBER/DIBER

        • Co-crystallization experiments of proteins with nucleic acids do not guarantee that both components are present in the crystal. While working as a PhD student with Matthias Bochtler, Grzegorz Chojnowski developed DIBER - a method with which to predict crystal content (DNA? protein? or both?) from the diffraction data. Now, we have together developed RIBER, which should be used when protein and RNA are in the crystallization drop. The combined RIBER/DIBER suite builds on machine learning techniques to make reliable, quantitative predictions of crystal content for non-expert users and high throughput crystallography. RIBER/DIBER requires diffraction data to at least 3.0 Å resolution in MTZ or CIF format.

      • MinkoFit3D

        • MinkoFit3D is a method for fitting macromolecular assemblies or their components into electron density maps. Our approach is based on finding “tight passages” inferred from the Minkowski sum boundary of two polyhedral surfaces of the structure and its map. Following the initial fit, either a robust brute-force or a genetic algorithm is used to build an initial assembly, and multi-body refinement is applied in direct electron density space.

      • Metadisorder

        • The GeneSilico MetaServer is a gateway to a number of third-party methods for protein structure prediction (identification of domains, secondary structure prediction, fold-recognition, and finally, 3D model generation). Users can submit a protein sequence (or alignment) by a single click, then analyze the summary of results generated by many methods, and finally predict the protein structure according to the "consensus" approach.

      • QA-RecombineIT

        • QA-RecombineIt provides a web interface to assess the quality of protein 3D structure models and to improve the accuracy of models by merging fragments of multiple input models. In the first stage (QA-mode), our server predicts the global quality of input models and provides estimates of local quality as the deviation between C-α atoms in the models and corresponding atoms in the unknown native structure. Together with the input models, these predictions subsequently become the input for the second stage (RecombineIt-mode), in which fragments predicted to be better than others are judiciously combined to generate hybrid (consensus) models. Finally, hybrid models are scored by the MQAPs implemented in the QA-mode and then presented to the user

      • NPDock

        • NPDock is a web server for modeling of protein-RNA and protein-DNA complex structures. It combines GRAMM for global macromolecular docking, scoring with a statistical potential, clustering of best-scored structures, and local refinement.

      • ClaRNA

        • ClaRNA is a new method for computational classification of contacts in RNA 3D structures. Unique features of the program are the ability to identify imperfect contacts and to process coarse-grained models. Each doublet of spatially close ribonucleotide residues in a query structure is compared to clusters of reference doublets obtained by analysis of a large number of experimentally determined RNA structures, and assigned a score that describes its similarity to one or more known types of contacts, including pairing, stacking, base–phosphate and base–ribose interactions.

      • BrickworX

        • BrickworX web server builds RNA and DNA models into electron density maps at resolutions as low as 4 Å. On input the program requires a MTZ file with structure factor amplitudes and phases. 

      • GDFuzz3D

        • GDFuzz3D is a method for protein 3D structure modeling, which starts from the protein sequence and a map of contacts between amino acid residues. It can handle predicted maps that contain e.g., probabilities of contacts between all residues, and a high fraction of incorrectly predicted contacts. It is not appropriate for predicting structures based on maps with only a few predicted contacts.

      • tRNAmodpred

        • The program takes as an input a set of unmodified tRNA sequences and a set of protein sequences corresponding to a proteome of a cell, identifies all RNA residues that correspond to known modification sites with known enzymes, finds homologs of known tRNA modification enzymes, and maps the predictions onto known pathways of RNA modification, to identify theoretically possible modification reactions for all positions in query tRNAs.

      • SupeRNAlign

        • SupeRNAlign is a tool for flexible superposition of RNA 3D structures. Our program implements an iterative algorithm that splits RNA structures into fragments and superimposes them using existing tools (currently R3D Align). Finally, the superimposed structures are saved to a .pdb file and a sequence alignment is generated.

    • TOOLKIT
  • Home
  • Funded grants
    • FNP (TEAM)

      • Modeling of dynamic interactions between RNA and small molecules and its practical applications (POIR.04.04.00-00-3CF0/16-00); 3 449 541 PLN; 2017-2020. PI: J.M.Bujnicki, vice-PI: F.Stefaniak

    • NCN (MAESTRO)

      • Integrative modeling and structure determination of macromolecular complexes comprising RNA and proteins (2017/26/A/NZ1/01083); 3 500 000 PLN; 2018-2023. PI: J.M.Bujnicki, vice-PI: N.Chandran

    • NCN (MINIATURA)

      • Photoswitchable ligands for riboswitches (2018/02/X/NZ1/01468); 21670 PLN 2018-2019. PI: F.Stefaniak

    • NCN (OPUS)

      • Development of new methods for designing RNA molecules that fold into desired spatial structures and their use for development of new functional RNAs and for prediction of noncoding RNAs in transcriptome sequences (2017/25/B/NZ2/01294); 1 494 250 PLN; 2018-2021. PI: J.M.Bujnicki, vice-PI: T.Wirecki

  • Achievements
  • Publications
  • In the media
  • People
  • Employment

Statistical potentials for RNA-Protein docking
 

Irina Tuszynska, Janusz M. Bujnicki
 
Docking methods are widely used to predict complexes of macromolecular structures. The biggest challenge of docking is to find near native structures among a set of alternative structures. Statistical potentials are used in methods for molecular interaction prediction. We provide two knowledge-based potentials for discrimination of near native protein-RNA decoys - DARS-RNP and QUASI-RNP. The QUASI-RNP potential uses quasichemical method to describe a reference state, while the DARS-RNP potential uses DARS (Decoys As Reference State).
 
Usage

Our knowledge-based potentials may be used on any Linux operating system. Simply download DARS-RNP potential or QUASI-RNP potential on your computer. Minimum requirements: Python version >= 2.6, the BioPython library version >=1.45 and the Numeric library.
 
Example:
Discrimination of native-like structures from unbound docking decoys of signal recognition particle (SRP) with RNA.
 
Unbound docking decoys were obtained by a low resolution docking with GRAMM of a signal recognition particle (SRP) (pdb code 1LNG, a complex with RNA) to RNA (pdb code 1Z43, apo form). As a reference we used the native structure of this complex (PBD code 1LNG). We produced 10000 decoys with the grid step = 3.0, and repulsion and attraction parameters 15 and 0, respectively. The grid-step radius was used as a projection of an atom. The systematic search through the rotational coordinates was performed every 15 degree.
 
We used QUASI-RNP potential to assess decoys and cluster 100 best scored structures.
 
python QUASI_potential_3.py –f list_10000.txt > QUASI_potential_3.out
 
Next we make the scatter plot of energy (from the output file) - RMSD (rmsd_file) dependence (Figure 1). The first biggest cluster consists of native-like structures (Figure 2).

 

Figure 1. An energy-RMSD dependence for 10000 decoys generated by GRAMM program for the complex of SRP with RNA. Cluster is colored red.


Figure 2. A native structure of SRP complex (shades of blue: dark blue - RNA, light blue - protein) and the best scored structure from the biggest cluster (shades of red: red - RNA, salmon pink - protein). RNA molecules are superimposed.