Frequently Asked Questions
What is SEEKR2?
Simulation Enabled Estimation of Kinetic Rates is, simultaneously, a collection of scientific software programs, and also a technical approach to estimate the kinetics and thermodynamics of certain molecular processes, especially ligand- receptor binding, using Milestoning theory. SEEKR2 is the second iteration of the software, with better speed, capabilities, and usability than the earlier version.
What can SEEKR2 do?
At present, the main utility of SEEKR2 is to estimate the kinetics of binding and unbinding (represented by the quantities k-on and k-off), in addition to the thermodynamics of binding (represented by the Gibbs free energy of binding).
Testing and preliminary calculations indicate that SEEKR2 can consistently obtain binding/unbinding rate constants within an order of magnitude of the experimental quantity, and sometimes much better, even for systems with very long residence times. SEEKR2 can also usually obtain Gibbs free energies of binding within ~1 kcal/mol of experiment. In addition, the SEEKR method has been able to rank compounds by their residence time and affinity to a receptor with good accuracy.
More details about SEEKR2’s accuracy and performance can be obtained from the publications on the Index page.
Development is ongoing, and more capabilities for SEEKR2 are being actively pursued. The kinetics and thermodynamics of intramolecular motion (such as hinge or pocket opening/closing), intersite transfer, peptide folding, and processes can be estimated using SEEKR2. Publications of these applications are forthcoming.
What if I doubt SEEKR2’s usefulness/validity?
How can I use SEEKR2 for my research?
How does one get a benchmark of a SEEKR2 calculation?
If the SEEKR2 run.py program terminates successfully, the last thing it will print is a benchmark (in ns/day) for an MD anchor. Therefore, if you wish to obtain a benchmark, consider running a short SEEKR2 calculation by passing a relatively small number (like 10000) to the “-t” argument of run.py.
How many CPUs are optimal for a SEEKR2 calculation?
In NAMD mode, using multiple CPUs is likely to make the calculation go faster. However, in OpenMM mode, the CUDA platform is designed to use only one CPU, so using multiple CPUs in OpenMM mode is not likely to make the calculation go any faster.