FAQs

We make every effort to secure your data and protect it. Our full privacy policy can be found on our Privacy Notice page.

While both water and fragment map data are generated with the same technique (grand canonical Monte Carlo simulation), the application is different.

Water map data is generally used to understand the cost of desolvation at sites on the surface of the protein. A water map will reveal areas where waters are easy or hard to displace and may provide hints about which direction to grow a candidate compound.

Fragment maps reveal where small chemical fragments have strong affinity on the surface of the protein. In BMaps, high-affinity fragments from simulation can be automatically substituted into a candidate compound.

Changes to affinity can be achieved by terminal group substitutions, scaffold hopping, and extending the compound into new pockets. Fragment binding data provides suggestions for each of these modifications, ranked by a free energy metric incorporating entropy considerations. BMaps provides a fragment search capability which interrogates the body of available fragment maps for fragments that bind in a configuration with a geometric relationship to the starting compound suitable for bond formation. Results are ranked by the free energy metric for prioritization.

When a compound binds in a given pose on the protein, the location and affinity of waters that have been displaced to achieve that pose provide an indication of what parts of the compound would benefit from substitution or not. Further, hydrogen bonding to waters which also bind to the protein can provide beneficial interactions that can be exploited. Finally, when looking to extend a molecule into a new sub-pocket, the presence of weakly bound waters versus tightly bound waters in that pocket provides a way to assess the potential of binding there. In BMaps, the bound waters are highlighted as weakly binding (green), moderate binding (gray), or strongly binding (red), and the quantitative ranking is seen by hovering the mouse over a particular water.

Water maps can be distinguished what information about water binding they provide. Maps using probing of the protein surface are inexpensive and can identify where individual waters are likely to bind but lack a rigorous free energy scoring to reliably rank the sites. Maps derived from molecular dynamics simulations can provide a reliable ranking of individual water binding sites by free energy. The BMaps water maps, derived from grand canonical Monte Carlo simulations, provide both the individual water free energy ranking and efficiently determine the lowest energy configurations of networks of water molecules (multi-body water interactions), and are significantly less expensive than other maps from simulations. The multi-body water configurations are particularly important for DNA/RNA and proteins that bind them. A new study comparing the various methods is being prepared for publication soon.

A list of our publications can be found on our Publications page.

A description of the simulation technology is on our Technology Overview page.

Deriving water maps, using grand canonical Monte Carlo simulations, has been simplified and mostly automated. After loading your structure, by selecting one of the installed BMaps structures or importing your own, press the “Modeling Tasks” button then “Run Water Sim.”. This dialog will allow you to configure the type of simulation you wish to do by selecting which molecule components to included in the simulation (apoprotein, protein-compound complex, etc.).

After configuring the simulation, click “Add to cart” to visit the simulation shopping cart and run the simulation.

A video tutorial can be found on our Quick Start Guide page.

Proteins installed in BMaps have been carefully prepared examining common issues in crystal structures such as missing atoms, modified amino acids or nucleotides, amide flips and histidine protonation states. Appropriate Amber forcefield parameters and bond orders have been assigned to all atoms and bonds. Over 100 fragments have been simulated on proteins installed in BMaps, including water with the apoprotein .

If the user uploads their own structures or downloads one from the PDB, an attempt will be made to add hydrogens and assign Amber forcefield parameters. While this is successful > 90% of the time, there are cases where errors will be reported, and manual preparation is required. Please contact Conifer Point if you require help with this. Note, the newest cryo-em structures, while exciting to look at, are not practical to simulate due to their large size.

When a compound is loaded into BMaps, either from a molecule file such as ChemDraw, or from the BMaps 2D editor, it needs to be placed in a suitable location relative to the protein. The default behavior is to align any new compounds with an existing compound or co-crystal ligand. This alignment is only between the old and new compounds and does not account for the presence of the protein. As such, the aligned position might result in overlaps with the protein. If small modifications are made in the 2D editor, an attempt is made to retain the 3D coordinates from the starting molecule, reducing the likelihood of protein overlaps. Usually, users will do an energy minimization following this alignment by clicking on “Calculate” button in the Energies tab of the information panel in the lower left of the BMaps workspace. If this doesn’t produce a good result, the next step is to dock the compound (click the “Dock” button under the “Modeling Tasks” button).

Note, if there are no existing compounds or co-crystal ligand present, the new compound is just placed nearby the protein, waiting for docking. If a compound is imported from a 2D format, such as SMILES or ChemDraw, a 3D conformation needs to be generated for it before alignment. This works well for compounds with a small number of rotatable bonds (say, 3-4), but does not work well for larger compounds with > 10 rotatable bonds, because the compounds can often curl up into conformations different from those seen in bound conformations. Docking may mitigate this, or a more appropriate 3D conformation of the compound needs to be generated outside of BMaps before importing it.

BMaps makes use of AWS (Amazon Web Services) to run a simulation management system. This facility queues requests for simulations and allocates cloud computing resources to run the simulations, one for each fragment to be simulated. The simulator reports the progress and completion of the jobs back to BMaps. This provides for a robust and economical computing platform. As many as 10,000 simultaneous simulations have been done in this way.

All BMaps service levels support the key computational features like docking and generating fragment maps. However, the monthly limits of the Basic service level (1 water map, 10 fragment maps, 100 docks) can be quickly exhausted in the course of full projects.

The Pro level subscription gives increased access to computation for docking and simulation operations and allows an extra BMaps session to be open. Note: new simulation data produced for Basic or Pro subscriptions may eventually end up in the public repository.

The Premier level includes even more computation, as well as isolated storage for simulation data. New simulation data produced for a Premier level subscription will always be kept private and will never be shared with the public repository.

See our Pricing page for more information.

The BMaps product was developed with support from the National Institute of General Medical Sciences (NIGMS) of the National Institutes of Health under award numbers R43GM109549 and R44GM109549. On-going costs of hosting on AWS are paid for by customer subscription revenue.

A raw fragment map is a set of statistical samples (snapshots) recorded during the GCMC fragment simulation. Typically 100 samples are taken for a given value of excess chemical potential (average free energy per fragment). To reduce the amount of data for searching fragment poses for bondable orientations, these samples are further subsampled by a factor of 10, and only fragments within 15A of defined hot-spots (active sites) are retained.  A fragment map summary, is a composite view of this data across all chemical potential values, where, at any given site on the surface of the protein where a fragment binds , the best affinity (lowest chemical potential) fragment pose is retained.