DelPhi DelPhi provides numerical solutions to the Poisson-Boltzmann equation (both linear and nonlinear form) for molecules of arbitrary shape and charge distribution. The current version is fast (the best relaxation parameter is estimated at run time), accurate (calculation of the electrostatic free energy is less dependent on the resolution of the lattice) and can handle extremely high lattice dimensions. It also includes flexible features for assigning different dielectric constants to different regions of space and treating systems containing mixed salt solutions. E.Alexov, R.Fine, M.K.Gilson, A.Nicholls, W.Rocchia, K.Sharp, and B. Honig. Finite Difference Poisson-Boltzman Solver Numerical Calculation of Electrostatic Potential http://trantor.bioc.columbia.edu/delphi Freely available to academia; pay model for commercial users. Fortran and C DelPhi takes as input a coordinate file format of a molecule or equivalent data for geometrical objects and/or charge distributions electrostatic potential in and around the system MAGNet Stable public release GRASP A molecular visualization and analysis program. It is particularly useful for the display and manipulation of the surfaces of molecules and their electrostatic properties. Anthony Nicholls and Barry Honig. molecular visualization Molecular Visualization Package http://trantor.bioc.columbia.edu/grasp Freely available to academia. Fortran PDB files, potential maps from DelPhi molecular graphics. MAGNet v1.3.6 .Stable public release. Nest Modeling protein structure based on a sequence-template alignment. The current server works only for modeling with a single template. Part of jackal, which can be downloaded. Xiang, Z. and Honig, B. modeling, protein structure, sequence-template alignment. Homology Modeling http://honiglab.cpmc.columbia.edu/cgi-bin/jackal/nest.cgi Freely available to academia. C++ pir and PDB files MAGNet Stable public release. JACKAL Jackal is a collection of programs designed for the modeling and analysis of protein structures. Its core program is a versatile homology modeling package nest. JACKAL has the following capabilities: 1) comparative modeling based on single, composite or multiple templates; 2) side-chain prediction; 3) modeling residue mutation, insertion or deletion; 4) loop prediction; 5) structure refinement; 6) reconstruction of protein missing atoms;7) reconstruction of protein missing residues; 8) prediction of hydrogen atoms; 9) fast calculation of solvent accessible surface area; 10) structure superimposition. Z. Xiang and B. Honig Protein Structure Modeling Prediction of Side-chain Conformations http://trantor.bioc.columbia.edu/programs/jackal Freely available to academia. C++ MAGNet Version: 1.5 as of Oct 20 2002, Stable public release. GRASP2 GRASP2 is an updated version of the GRASP program used for macromolecular structure and surface visualization, contains a large number of new features and scientific tools: Enhanced GUI; Structure alignment and domain database scanning; A gaussian surface generator and new surface coloring schemes; Sequence visualization and alignment; Completed work can be stored in "project files; Among the many objects that can be stored in a project file are views of the structure; defined subsets, surfaces; Direct printing to printers at full printer resolution. Donald Petrey and Barry Honig. molecular visualization Molecular Visualization Package http://trantor.bioc.columbia.edu/grasp2 Freely available to academia. C++ PDB files, potential maps from DelPhi, sequence alignments. molecular graphics, structural alignments. MAGNet Stable public release PrISM PrISM is an integrated computational system where computational tools are implemented for protein sequence and structure analysis and modeling. Wang, L, Yang, A. S. & Honig, B. protein analysis/modeling Homology Modeling http://trantor.bioc.columbia.edu/programs/PrISM/ Freely available to academia. Fortran MAGNet Stable public release Protein-DNA interface alignment The protein-DNA alignment software allows one to align the interfacial amino acids from two protein-DNA complexes based on the geometric relationship of each amino acid to its local DNA. Siggers, T.W., Silkov, A & Honig, B. protein-DNA interface Prediction of Side-chain Conformations http://trantor.bioc.columbia.edu/programs/intfc_aln Freely available to academia C++ and Perl two PDB files that both contain protein-DNA complexes The programs will output the aligned residues and their corresponding residue-residue similarity scores, s(i,j). MAGNet Stable public release. SURFace SURFace algorithms are programs that calculate solvent accessible surface area and curvature corrected solvent accessible surface area Nicholls, A., Sharp, K., Sridharan, S. and Honig, B. solvent accessible surface area Caculation of Solvent Accessible Area http://trantor.bioc.columbia.edu/surf/ Freely available to academia. MAGNet Stable public release. Target Explorer Automated process of prediction of complex regulatory elements for specified set of transcription factors in Drosophila melanogaster genome. Target Explorer is a complex tool with user-friendly self-explanatory Web-interface that allows to user: 1. create customized library of TF binding site matrices based on user defined sets of training sequences; 2. search for new clusters of binding sites for specified set of TFs; 3.extract annotation for potential target genes. Sosinsky A, Bonin CP, Mann RS, Honig B. prediction of binding sites for transcription factors Sequence Annotation http://trantor.bioc.columbia.edu/Target_Explorer/ Freely available to academia. perl, cgi genomic sequences clusters of known binding sites MAGNet Stable public release. MEDUSA and Gorgon MEDUSA is an algorithm for learning predictive models of transcriptional gene regulation from gene expression and promoter sequence data. By using a statistical learning approach based on boosting, MEDUSA learns cis regulatory motifs, condition-specific regulators, and regulatory programs that predict the differential expression of target genes. The regulatory program is specified as an alternating decision tree (ADT). The Java implementation of MEDUSA allows a number of visualizations of the regulatory program and other inferred regulatory information, implemented in the accompanying Gorgon tool, including hits of significant and condition-specific motifs along the promoter sequences of target genes and regulatory network figures viewable in Cytoscape. David Quigley, Manuel Middendorf, Steve Lianoglou, Anshul Kundaje, Yoav Freund, Chris Wiggins, Christina Leslie Signaling network reconstruction http://www.cs.columbia.edu/compbio/medusa (MATLAB),http://compbio.sytes.net:8090/medusa (Java beta version) Open source Java (prototyped in MATLAB) Discretized (up/down/baseline) gene expression data in plain text format, promoter sequences in FASTA format, list of candidate transcriptional regulators and signal transducers in plain text format. Regulatory program represented as a Java serialized object file readable by Gorgon and as a human readable XML file. Gorgon currently generates views of learned PSSMs, positional hits along promoter sequences, and views of the ADT as HTML files, and generates network figures as Cytoscape format files. MAGNet Version 2.0 July 2006 pre-release beta version; Version 1.0 (MATLAB), April 2005, stable public release String kernel package The string kernel package contains implementations for the mismatch and profile string kernels for use with support vector machine (SVM) classifiers for protein sequence classification. Both kernels compute similarity between protein sequences based on common occurrences of k-length subsequences ("k-mers") counted with substitutions. Kernel functions for protein sequence data enable the training of SVMs for a range of prediction problems, in particular protein structural class prediction and remote homology detection. A version of the Spider MATLAB machine learning package is also bundled with the code, which allows users to train SVMs and evaluate performance on test sets with the packaged software. Eleazar Eskin, Rui Kuang, Eugene Ie, Ke Wang, Jason Weston, Bill Noble, Christina Leslie Protein Modeling and Classification http://www.cs.columbia.edu/compbio/string-kernels Open source String kernel code is implemented in C. Spider is a set of object-oriented MATLAB routines. The mismatch kernel requires sequence data in FASTA format. The profile string kernel uses probabilistic profiles, such as those produced by PSI-BLAST, in place of the original sequences. The Spider SVM implementation requires both the kernel matrix and a label file of binary or multi-class labels for the training data; this data must be loaded into MATLAB variables before using Spider routing. The kernel code produces a kernel matrix for the input data in tab-delimited text format. The Spider package trains SVMs and stores the learns classifier and results from applying the classifier on test data as MATLAB objects. MAGNet Version 1.2 September 2004 stable public release MatrixREDUCE Regulation of gene expression by a transcription factor requires physical interaction between the factor and the DNA, which can be described by astatistical mechanical model. Based on this model, the MatrixREDUCE algorithm uses genome-wide occupancy data for a transcription factor (e.g.ChIP-chip or mRNA expression data) and associated nucleotide sequences to discover the sequence-specific binding affinity of the transcription factor. The sequence specificity of the transcription factor's DNA-binding domain is modeled using a position-specific affinity matrix (PSAM), representing the change in the binding affinity (Kd) whenever a specific position within a reference binding sequence is mutated. The PSAM can be transformed into affinity logo for visualization using the utility program AffinityLogo, and a MatrixREDUCE run can be summarized in an easy-to-navigate webpage using HTMLSummary. Barrett Foat, Xiang-Jun Lu, Harmen J. Bussemaker position-specific affinity matrix, binding affinity, cis-regulatory element, expression data, ChIP-chip, transcription factor Signaling network reconstruction http://www.bussemakerlab.org/software/MatrixREDUCE ANSI C, making use of Numerical Recipes routines. sequence file in FASTA format; and expression data file in tab-delimited text format. PSAMs in numeric and graphical format, parameters of the fitted model, and an HTML summary page. MAGNet Version 1.0 July 10 2006, extensively tested in lab. T-profiler T-profiler is a web-based tool that uses the t-test to score changes in the average activity of pre-defined groups of genes. The gene groups are defined based on Gene Ontology categorization, ChIP-chip experiments, upstream matches to a consensus transcription factor binding motif, and location on the same chromosome, respectively. If desired, an iterative procedure can be used to select a single, optimal representative from sets of overlapping gene groups. A jack-knife procedure is used to make calculations more robust against outliers. T-profiler makes it possible to interpret microarray data in a way that is both intuitive and statistically rigorous, without the need to combine experiments or choose parameters. André Boorsma, Barrett C. Foat, Daniel Vis, Frans Klis, Harmen J. Bussemaker gene expression, transcriptome, ChIP-chip, Gene Ontology Network characterization http://www.t-profiler.org T-profiler is written in PHP, data is managed by a MYSQL database server Currently, gene expression data from Saccharomyces cerevisiae and Candida albicans are supported. MAGNet TranscriptionDetector A tool for finding probes measuring significantly expressed loci in a genomic array experiment. Given expression data from some tiling array experiment, TranscriptionDetector decides the likelihood that a probe is detecting transcription from the locus in which it resides. Probabilities are assigned by making use of a background signal intensity distribution from a set of negative control probes. This tool is useful for the functional annotation of genomes as it allows for the discovery of novel transcriptional units independently of any genomic annotation. Xiang-Jun Lu, Gabor Halasz, Marinus F. van Batenburg tiling arrays, expression, transcriptome http://www.bussemakerlab.org/software/TranscriptionDetector/ ANSI C, making use of GSL. Expression data (GEO or other platforms) and designation of which probes represent negative controls and which are data probes. A text file with a list of probes corresponding to significantly expressed loci. MAGNet PhenoGO PhenoGO adds phenotypic contextual information to existing associations between gene products and Gene Ontology (GO) terms as specified in GO Annotations (GOA). PhenoGO utilizes an existing Natural Language Processing (NLP) system, called BioMedLEE, an existing knowledge-based phenotype organizer system (PhenOS) in conjunction with MeSH indexing and established biomedical ontologies. The system also encodes the context to identifiers that are associated in different biomedical ontologies, including the UMLS, Cell Ontology, Mouse Anatomy, NCBI taxonomy, GO, and Mammalian Phenotype Ontology. In addition, PhenoGO was evaluated for coding of anatomical and cellular information and assigning the coded phenotypes to the correct GOA; results obtained show that PhenoGO has a precision of 91% and recall of 92%, demonstrating that the PhenoGO NLP system can accurately encode a large number of anatomical and cellular ontologies to GO annotations. The PhenoGO Database may be accessed at www.phenogo.org. Yves Lussier and Carol Friedman are the principal investigators. The programmers are Jianrong Li, Lee Sam, and Tara Borlawsky Phenotypic integration, computational phenotypes Natural Language Processing http://www.phenogo.org n/a A variety of modules, the web portal is in Java and MySQL, the computational terminology component (phenOS) is written in Perl scripts that queries tables in IBM DB2, the natural language processing component is written in PROLOG. Gene Ontology Annotations Files and Medline Abstracts XML file and www.phenogo.org Web Portal MAGNet Version 2 Feb 2006 MINDY Given a transcription factor of interest, MINDY uses a large set of gene expression profile data to identify potential post-transcriptional modulators of the transcription factor's activity. MINDY is based on a three-way statistical interaction model that captures the post-transcriptional regulatory event where the ability of a transcription factor to activate/repress its target genes is monotonically controlled by a potential modulator gene. Kai Wang, Ilya Nemenman, Adam Margolin, Riccardo Dalla-Favera, Andrea Califano gene expression, transcriptional interaction, modulator Signaling network reconstruction n/a n/a C++ and MATLAB, Java Gene expression data in the EXP format, and a user-specified transcription factor of interest Lists of the putative modulators and target genes of the transcription factor, and the modulatory interactions involving them MAGNet Stable release April 2007 B Cell Interactome The B cell interactome (BCI) is a network of protein-protein, protein-DNA and modulatory interactions in human B cells. The network contains known interactions (reported in public databases) and predicted interactions by a Bayesian evidence integration framework which integrates a variety of generic and context specific experimental clues about protein-protein and protein-DNA interactions - such as a large collection of B cell expression profiles - with inferences from different reverse engineering algorithms, such as GeneWays and ARACNE. Modulatory interactions are predicted by MINDY, an algorithm for the prediction of modulators of transcriptional interactions. Lefebvre C, Lim WK, Basso K, Dalla Favera R, and Califano A. Naive Bayes, Mixed-Interaction Network, human B cells. Interaction Modeling http://amdec-bioinfo.cu-genome.org/html/BCellInteractome.html Perl n/a text file of binary interations associated with a probability. MAGNet Version 2 March 2007 ARACNE ARACNE is an algorithm for inferring gene regulatory networks from a set of microarray experiments. The method uses mutual information to identify genes that are co-expressed and then applies the data processing inequality to filter out interactions that are likely to be indirect. Margolin AA, Nemenman I, Basso K, Wiggins C, Stolovitzky G, Dalla Favera R, Califano A. Reverse engineering, mutual information, genetic networks, microarray Signaling network reconstruction http://amdec-bioinfo.cu-genome.org/html/ARACNE.htm Open source C++, Java Text file containing measurements from a set of microarray experiments. Text file containing predicted interactions. MAGNet Version 1 June 2006 geWorkbench geWorkbench is a Java application that provides users with an integrated suite of genomics tools. It is built on an open-source, extensible architecture that promotes interoperability and simplifies the development of new as well as the incorporation of pre-existing components. The resulting system provides seamless access to a multitude of both local and remote data and computational services through an integrated environment that offers a unified user experience. Over 50 data analysis and visualization components have been developed for the framework, covering a wide range of genomics domains including gene expression, sequence, structure and network data. A. Califano, A. Floratos. M. Kustagi, K. Smith, J. Watkinson, M. Hall, K. Keshav, X. Zhang, K. Kushal, B. Jagla, E. Daly, M. VanGinhoven, P. Morozov. Analysis suite, gene expression analysis, sequence analysis, network reconstruction, structure predcition, visualization. Visualization http://www.geworkbench.org Free. Java Gene epxression data (Affy, GenPix, RMA), Sequence (FASTA), Structure (PDB). Analysis results (multiple formats). MAGNet 1.0.5 3/23/07 stable production release