Torch V10.4 released

The latest version of Torch, our 3D molecule design tool for medicinal chemists and synthetic chemists is now available. Torch uses Cresset’s description of a molecule’s shape and electrostatic character to help you understand the root causes of activity and use this to design new compounds. V10.4 has new functionality to aid decision making around new designs and an enhanced user interface.


  • New Multi-parameter Optimization (MPO) options including condensing many activity and physicochemical properties into a single score representing the fit to the project profile
  • Color molecules and table cells using the MPO score (Figure 1)
  • Color molecules in the protein binding site using activity or logP or other physicochemical properties
  • View molecules as Tiles with selected data
  • Divide your dataset into different Roles and add tags to label important compounds or subsets.

Figure 1: The colorful new Tiles view and histograms in TorchV10.4


Torch V10.4 has hundreds of improvements on the previous versions that will enhance your experience. If you are currently a user then get in touch for download links. If you haven’t tried Torch before get your free evaluation.

Multi-parameter optimization in Torch


Figure 2: Enhanced Radial Plot

In Torch V10.3 the Radial Plot was introduced to provide a graphical representation of numerical data. In particular it created a simple picture to show how a molecule fits the physicochemical profile of a project with the idea that parameters are within an ideal range, an unacceptable range or somewhere in between. V10.4 enhances this representation by combining all the scores in the Radial Plot together into a single number scaled between zero and 1 (Figure 2).

Thus compounds with a radial plot score of 1 fit the project profile perfectly while those with zero lie outside the desirable property space in all aspects. Since not all properties are equally important, Torch enables a weighting factor that can be applied to each property (Figure 3). The weight is used to scale the contribution to the final score. This is useful when you want to focus on one property more than another, for example you are prepared to have a non-ideal value for MW if the logP and TPSA are within the ideal range or you want a visual representation of that property but not have it count towards the score.

Figure 3: The configuration of the Radial Plot now includes a weight to apply to each property in combining the properties into a single score.

Coloring of molecules in the 3D window now has options to apply a color scale based on any numerical property in the molecule table including the new radial plot score (Figure 4). This simple feature requires you to define how the scale should be applied to the numbers (by means of the Radial Plot property configuration). Once this is complete you are able to investigate how the 3D alignment and shape of molecules varies with the individual physicochemical properties and the overall project goals.


Figure 4: Nine DPP4 active molecules colored by activity (left) and Radial Plot Score (right). Green represents a better fit to the ideal values for the two properties (activity, radial plot score). In this case the most active compounds (top row) do not fit the overall project profile as well as those with lower activity (bottom row).


Helping SAR interpretation – molecule organization, tagging and plotting

This release enhances Torch’s ability to study SAR in multiple ways. You can now divide molecule datasets into separate sections, tag any molecule with an arbitrary text based label benefit from improved plotting capabilities to study numerical data and select molecules.

Molecule roles were introduced in the first release of Forge to enable it’s 3D-QSAR functionality. However, we found that we use this functionality extensively even when we are not interested in 3D-QSAR, for example, to group compounds into series or split out those that are designs from those that have been tested or are out for synthesis. We use molecule Roles (Figure 5) to help cluster datasets using 3D similarity with Activity Miner, to look at different sub-groups, as well as for the original purpose – 3D-QSAR. In bringing this functionality to Torch we are sure you will find it equally useful.

Figure 5: Tiles view of molecules split by Roles showing Radial Plot, Notes, and Tags on each medium sized tile.
Molecule Roles enable the categorization of molecules into single bins. To complement this, molecule Tags have been introduced to enable the labeling of molecules across multiple roles with any number of arbitrary labels. We tag compound sources, stereochemistry, designed vs tested, personal favorites, R-group variations and different series. When combined with the ability to filter the molecules based on these tags and the new ability to filter on substructures you get excellent flexibility to investigate even large SAR datasets.

Plotting histograms and scatter plots is now possible with Torch V10.4 (Figure 6). This simple enhancement comes with interactive selection of molecules and significantly improves the study of molecule properties across a dataset or within a role. For example, selecting a molecule in a scatter plot of Activity vs logP highlights the selection in a histogram of MW. Combined with the display of the selected in the 3D window it provides added flexibility in SAR analysis.

Figure 6: Scatter plots and Histograms in Torch V10.4 are interactive, selecting in one graph highlights the same molecules in other graphs.

Tiles view

To complement Tags and Molecules Roles we have introduced the Tiles view of molecules that has been so popular in Spark. This view presents any of the data that is in Torch on a tile and enables you to focus on the information that is important to you at the same time as seeing as much of the dataset as you can.

New models of activity

In Forge, Activity Atlas models have been introduced as a new way of summarizing your SAR in a simple yet detailed 3D view (Figure 7). In Torch these models act as templates to design new molecules. They are present in the 3D window and also in the editor to enable the interactive design (in 3D) of molecules that explore new regions of SAR or help you focus on maintaining activity while perfecting your macro-properties. We are confident that Activity Atlas models will transform your view of molecule design.

Figure 7: Activity Cliff Summary for Adenosine A1 (left); Adenosine A2a (middle); and Adenosine A3 (right).
Color coding of maps: Red: more positive increases activity; Blue: more negative increases activity; Green: steric bulk in this position is favorable; Magenta: steric bulk in this position is detrimental.


Try Torch V10.4

To try the new features in Torch V10.4 simply download a free evaluation or contact us for download links.

Try Cresset solutions on your project

Request a free software evaluation