Cresset command line binaries

Cresset supplies command line versions of our ligand based design workbench Forge and of our bioisostere tool Spark as well as for our conformation hunting algorithm XedeX and ligand minimizer XedMin. All these tools are available directly from the command line or through KNIME or Pipeline Pilot interfaces. Lastly there are a number command line scripts that interface to our Blaze virtual screening system to perform standard tasks like collection upload or searching. These scripts are largely replaced by with new REST api for Blaze.

The individual command lines are detailed below. Full details for operating the binaries are available through the relevant ‘man’ pages on Linux systems and html files on other operating systems.

Forge command line components


The Forge alignment algorithm can be accessed from using this command line binary. This binary reads molecules to be aligned and a reference structure or ‘template’ and outputs the molecules aligned to the reference. Molecules to be aligned can be read as conformation populations or as flat files (or smiles strings) which are conformationally explored (using an in-built version of XedeX) as part of the process.


The Activity Miner command line binary takes aligned molecules (such as those from falign) and outputs either a similarity or a disparity matrix. It is used to identify regions of structure activity relationship where small changes in structure result in high changes in activity or to cluster molecules using 3D metrics.


The Forge build binary takes aligned structures (such as those from falign) and constructs a 3D-QSAR model to explain the observed activity. It determines a set of ‘sample positions’ around the molecules based on their field points, and then calculates the value of the molecular fields at those positions. The data matrix derived from this is then processed by partial least squares (PLS) to generate a linear model for activity. The input molecules should be read in SDF format. A number of files are output to the current working directory including a Forge project file containing the model, the sample data (in CSV format) and all input molecules with a tag added for predicted activity.


The Forge score binary uses a designated Forge model to predict activities for molecules that have been previously aligned (e.g. using falign). Molecules are read in SDF, mol2 or XED format and output in SDF format to the STDOUT channel.


ftemplater is the command line equivalent to the FieldTemplater module of Forge. Ftemplater creates 3D models of binding (‘templates’, or ‘pharmacophores’) from cross comparison of diverse chemcial structures. Given three different structures that bind to the same active site in 2D, ftemplater generates a set of 3D models that detail how these molecules align in 3D. These ‘templates’ can be used as reference structures for ligand alignment (using falign), bioisosteric searches (using sparkcl) or virtual screening (using Blaze).


This simple binary is used to exact information and molecules from Forge (and Torch) project files. It is primarily useful within KNIME and Pipeline pilot as a link to other applications.

Spark command line components


This binary reads in a molecule, removes a user-specified section, and then searches for bioisosteric replacements for that section in a database of fragments. The replacement fragments are merged into the original molecule, and the result is scored by field and shape similarity to the original or other molecules. The portion to be replaced is specified using a set of atom ids or bond ids and options enable control of the type of atom that is used to form any new bond. The results are sent to the STDOUT channel as a SDF, mol2 or XED file. The binary requires access to databases of fragments that are supplied by Cresset or can be calculated using the sparkdb.exe binary.


Fragment databases for use by the GUI or command line Spark application can be created with this binary. Databases can be created by the fragmentation of whole molecule, the import of existing fragments or the processing of chemical reagents with an optional conformation hunt being performed on all fragments before storing in the database. Options control most of the parameters for the calculation.

Generic components


The calculation engine for GUI and command line binaries. FieldEngine is started directly by the primary interface to perform specific calculation tasks and hence no knowledge of its workings are required in normal operation. By default, one FieldEngine binary is started for each cpu core on the host machine (but this can be reconfigured in the application preferences or by command line switches). The applications communicate with the FieldEngine using a TCP/IP protocol that enable deployment of FieldEngines to remote computing resources.