News
Version 9 of Flare, Cresset's CADD solution, brings new and enhanced scientific features and methods for all users. These include Homology Modeling to create reliable 3D models for protein targets for which crystallographic information is not available, ensemble Molecular Mechanics/Generalized Born Surface Area (MM/GBSA) on dynamics trajectories for estimating binding free energy, a full integration of Spark in Flare, and enhanced protein preparation.
In this release, we have also further expanded and enhanced the choice of visual tools to investigate the results of Flare FEP experiments and analyze dynamics trajectories, and added many new options for fine-tuning the experiments.
Furthermore, the enhanced Extension Manager provides a streamlined workflow for the installation of Flare Python Extensions.
Experimental information about the 3D protein structure may not be available for many interesting biological targets, for example, many GPCRs, ion channels and novel targets. In all these cases, Homology Modeling can be used to create a reliable 3D structure to use in structure-based studies, from docking and scoring to molecular dynamics.
Homology Modeling in Flare uses the robust ProMod31 algorithm, deployed on a secure server hosted by Cresset. This enables Flare to build reliable homology models in just a few minutes, starting from aligned target sequences and template chains with sequence identity as low as 30%.
Figure 1. To build a homology model, Flare connects to a secure server hosted by Cresset. The model is built starting from aligned target sequences (A1_Target_Sequence) and template chains (8HDP_L).
Flare’s flexible implementation allows users to create both single-chain and multi-chain models (Figure 2). Furthermore, it is possible to include ligands and cofactors in the Homology Modeling experiment: this typically results in more realistic conformations for the side chains of the residues lining the binding site.
Figure 2. A known A2BR X-ray structure (PDB: 7LD4, left) was used as a template for modeling the adenosine A1 receptor (right) in a multi-chain homology modeling experiment.
MM/GBSA2 is a popular method to estimate the binding free energy for ligand-protein complexes, providing a good balance between accuracy of predictions and computational efficiency.
Building on the implementation of the method in Flare V8, this latest release expands the MM/GBSA capabilities of Flare to include the option to run the calculations on an ensemble of conformations from a molecular dynamics experiment. MM/GBSA Dynamics in Flare allows you to score multiple ligand-protein ensembles at a time, choosing from a choice of different implicit solvents, and selecting the desired trajectory range and score interval (Figure 3).
Figure 3. MM/GBSA calculations in Flare can be run on individual, minimized ligand-protein complexes (MM/GBSA Single Point) as well as on pre-calculated Dynamics trajectories (MM/GBSA Dynamics).
This typically leads to a more accurate estimate of the binding free energy with respect to single-point MM/GBSA calculations (performed on a single minimized protein-ligand complex), as shown in Figure 4.
Figure 4. Results of MM/GBSA single point (left) and MM/GBSA Dynamics (right) experiments on the JNK1 data set from Wang et al.3
Flare V9 features a full integration of Spark, Cresset’s best-in-class bioisostere application.4 This is the result of a two-year project, during which we have worked hard to give Spark users access to the variety of methods and features which Flare provides for preparing Spark experiments and optimizing results, at the same time carefully preserving the ease of use of Spark. Using Spark in Flare offers several advantages, giving you access to:
We carefully transferred the Spark wizards, the heart of Spark’s minimal learning curve, the user-friendly interface for creating and updating Spark fragment databases, and the clustering algorithm, into the dedicated tab which you can see in Figure 5.
Figure 5. The Spark tab in Flare.
We have streamlined these Spark wizards, an example of which is the 'Scaffold Hopping or R-group replacement' wizard shown in Figure 6. In the first panel (Figure 6 – left), circle the core of the molecule you want to replace. In the second panel (Figure 6 – middle), use the attachment point chooser (red box) to move across the different attachment points, and set the allowed atom types. The 'Reference & Protein' panel (Figure 6 – right) enables you to choose any reference molecules you want to use to guide the scaffold hopping experiment (these can be in the same role as the starter molecule, or in a different role) and whether to use a protein as an excluded volume or for docking.
Figure 6. The 'Scaffold Hopping or R-group replacement' wizard in Flare.
Spark provides access to an outstanding collection of millions of bioisosteres to generate new project ideas. In the enhanced 'Database selection & Advanced options' panel (Figure 7) you can select the databases you wish to search, set an experiment name and a role for the results, and choose whether to remove duplicate results from different Spark experiments.
Overall, a scaffold-hopping experiment can be started in just 4 clicks.
Figure 7. In the enhanced 'Database selection & Advanced options' panel you can define an experiment name (which will be used to name the results, and optionally also can be used as a Tag), define a role for saving the results, and choose to remove duplicate Spark results from different Spark experiments saved into the same role.
One of the many advantages of running Spark experiments in Flare is that you are no longer limited to one starter molecule, one protein, one set of results (in practice, one Spark experiment) per project. Flare enables you to run multiple different experiments and keep the results nicely organized in different roles within the same Flare project.
Furthermore, all Spark features are now accessible from the Flare Python API. This makes it easy to set up custom workflows (for example, run a Spark search and automatically re-dock results, or automatically score them with MM/GBSA). Command-line scripts are also available as a drop-in replacement for existing Spark command-line tools.
Several new features and analysis tools for molecular dynamics are available in this release. Below we will highlight some of the most important new additions.
We have added the Principal Component Analysis (PCA) plot (Figure 8) to the wide and growing choice of analysis tools which help Flare users analyze and understand Dynamics results. You can use this visual tool to monitor the most important motions of your biomolecule over the Dynamics trajectory.
Starting from the coordinates of each heavy atom along the trajectory, the method follows the protein dynamics along its Principal Components (PCs), representing the directions of greatest variance in the data, with the first few PCs capturing the most significant modes of motion in the system.
In a recent article, we explore how this visualization tool helps to reduce the dimensionality of the dataset; revealing hidden trends and highlighting significant information for effective analysis, driving faster project outcomes.
Figure 8. Use the new PCA plot in Flare to monitor the most important motions of your biomolecule over the Dynamics trajectory.
If you wish to run multiple molecular dynamics experiments for different ligands within the same protein, in Flare V9 you can automate the process. Select the desired ligands and protein, choose the options for running the experiment, and press the 'Start' button (Figure 9). Flare will run separate Dynamics experiments for each ligand-protein complex, and the resulting trajectories will be appended to the project.
Figure 9. In Flare V9 you can start multiple Dynamics experiments for the selected ligands and a chosen protein in one go.
We have also added several new advanced options which you can use to fine-tune your experiments. For example, you can now create, use and save bespoke equilibration protocols, by adding/removing existing steps from the 'Standard' protocol, and by changing the conditions for performing each step, as shown in Figure 10.
Figure 10. Create, use and save bespoke custom equilibration protocols for Dynamics and Flare FEP experiments.
We have also added several additional lipid membrane models for performing molecular dynamics on membrane proteins, and added the possibility to select the explicit water model to use with each model.
Finally, among many other new nice features, you can now:
Flare FEP includes several new and enhanced features in Flare V9, with some of the most interesting changes summarized below.
An enhanced Activity Plot enables you to monitor the precision of the FEP calculation, expressed as the proportion of the predicted active molecules which are indeed experimentally active. The desired activity threshold for separating active/inactive molecules can be set using the slider at the top of the graph (Figure 11). Furthermore, in Flare V9 you can change the activity unit for drawing the plot, switching from the default ΔG to µM or nM activity.
Figure 11. Monitor the precision of Flare FEP calculations by defining an activity threshold using the slider.
The new graph error analysis is an additional tool which helps you to troubleshoot FEP results, by color-coding all the calculated links in the perturbation network based on the contribution they make to the overall calculation error (Figure 12). This facilitates quick and easy identification of problematic links which have a detrimental effect on the precision of the Flare FEP calculation.
Figure 12. The graph error analysis colors each link according to the contribution it makes to the overall error of the Flare FEP calculation. Blue: low contribution, grey: average contribution, red: high contribution. The link at the bottom is clearly problematic, as it is affected by hysteresis and contributes significantly to the overall error.
The FEP calculation dialog (Figure 13) has been enriched with several new options for fine-tuning the calculation.
Figure 13. New options in the Flare FEP calculation dialog.
The new 'Adaptive Lambda Schedules' option (Figure 13 – left) can be used to balance speed vs. accuracy for the adaptive lambda schedule. Choosing 'Fast' generates less lambda windows and accordingly leads to faster calculations, but with a higher chance of poor overlap. 'Accurate' generates more lambda windows, lowering the chance of poor overlap at the cost of longer calculations. 'Normal' uses the default algorithm as in previous versions of Flare FEP.
Another new option (Figure 13 - right) allows to set different simulation lengths for the free and bound state of the transformation.
As with Dynamics, several additional lipid membrane models are available for FEP calculations on membrane proteins (Figure 13 – right), with the possibility to select the explicit water model to use with each model; and it is possible to create and use bespoke equilibration protocols also for Flare FEP, as in Figure 10.
Finally, the 'Start' button now offers the option to recalculate only the ΔΔGs when calculations are complete (useful for example when molecules are removed from the perturbation network).
In this release of Flare, we have introduced new options to the 'Protein Preparation' dialog, to make the preparation of the biomolecular system under study more flexible and comprehensive.
New options are available to:
A new 'Create Biological Assemblies' button in the Protein tab also enables the creation of biological assemblies separately from protein preparation, as shown in Figure 14.
Figure 14. Protein preparation in Flare V9 features several new options. 'Remove post-translational modifications' was used to remove glycosylation from the protein chains of PDB: 6BBD (left) after recreating the biological assembly with the new 'Create Biological Assemblies' button in the Protein tab. The prepared biological assembly is shown to the right.
An enhanced Extension Manager (Figure 15) enables the seamless installation and update of Cresset or customer-created Python extensions with minimal user intervention. Extension packages can be installed manually by end users, or programmatically by IT, and it is possible to decide whether to update the extensions manually or automatically, whenever a new or updated extension package is available.
Figure 15. Seamless installation and update of Cresset or customer-created Python extensions with minimal user intervention with the enhanced Extension Manager in Flare V9.
In Flare V9 we have implemented a time-saving usability improvement to help you quickly navigate to recent files and folders. The File menu shows, in addition to the list of recent projects, a list of files you can select with a click (Figure 16) to easily re-open them, and frequently visited places to navigate to.
Figure 16. Easily reopen Flare projects and molecule files, and navigate to the desired folders and locations.
At every release of Flare, we work very hard at implementing not only additional interesting science, but as many as possible usability enhancements, many of which are in direct response to feedback from our users. Some other interesting new features are listed below:
And many others, which you can find detailed in the Flare release notes.
Flare V9 brings powerful new scientific methods, analysis tools and usability enhancements that give you a detailed understanding of your ligand-protein complexes.
Try Flare today to see for yourself how the rich and user-friendly interface gives you the insights you need to progress lead optimization with confidence.
Get in touch to arrange an evaluation and quickly get started in accessing Flare's wide range of features. Our dedicated team are on hand to support you during installation and set-up and you can access our vast range of tutorials introducing you to our common workflows, through to more advanced methods and functionality. We’re here to help you reach your goals, faster, enabling you to make the molecules that matter.
Abstract Pyrethroids are a common class of pesticide which are used in crop and personal protection against insects. The insecticidal ...
We present the results from a free energy perturbation calculation demonstrating a charge change between ligands in Flare FEP.
Overview In drug discovery, detailed knowledge of protein-ligand interactions are key to understand the recognition processes between the two molecules, ...
