Home

Optimal Probes is a software package to predict the smallest set of residue pairs for site-directed spin labeling for DEER/EPR experiments that best capture the slowest dynamics in a protein of interest. It uses molecular dynamics (MD) simulation datasets and a hyperparameter optimization strategy¹ for a Markov state model (MSM).

The MD simulation data is a conformational dynamics resource which can be exploited to extract residue pair distances which are important to study the slow processes in the protein’s dynamics as well as distances which provide insight into distinct conformational states of the protein. These residue pairs which are involved in the slow timescale processes of the protein, once determined from the simulation datasets, can then be used to perform actual EPR/DEER experiments. You can also make use of biased MD simulation data, obtained from using accelerated MD, steered MD, umbrella sampling, metadynamics or replica exchange techniques which can provide thermodynamic observables.

This website explains the workflow and is a guide to the user to obtain their own predictions.

For more information, please visit http://www.shuklagroup.org.

Citation

Mittal S and Shukla D. "Optimal Probes: An Efficient Method to Select Deer Distance Restraints using Machine Learning". Biophysical Journal 112.3 (2017).

Mittal S and Shukla D. "An Algorithm to Predict Optimal EPR/DEER Label Positions to Study Protein Dynamics". Submitted (2017)

Selvam B, Mittal S and Shukla D. "In silico Predictions of Conformational States of Bacterial Transporter PepT_So". Submitted (2017)

References

[1] McGibbon, R T and Pande, V S. "Variational cross-validation of slow dynamical modes in molecular kinetics." The Journal of chemical physics 142.12 (2015).