i submitted my article in where i conducted molecular dynamic simulation 100 ns for a complex protein-ligand. Reviewer asked us for more work to be done for this study as below
''More extensive simulations of longer simulations could make this article more appropriate. This could be done with structural visualization of clusters from non-parametric clustering or of structures from free-energy wells on PCA-based free-energy surfaces. And then significant discussion of the result protein structures and conformational changes".
can you please explain to me what he mean by non-parametric clustering and share with some articles examples that include example of what reviewer expect from us.
thannk you.
Non-parametric clustering is usually graph-based, where edges in the graph are somehow defined from the samples. In MD, if the order parameter is not mentioned, assume it is RMSD. So maybe he is asking you to build a graph based on RMSD, then prune the graph (min-cut algorithm) to get clusters. But the reviewer is being vague. With regard to the simulation length, you can prove that your simulation is representative by splitting it in the middle and doing the analysis on the two haves separately. Do they give the same answer? If they do then 100ns was enough. But don't do what I say, be smart about it. Do the analysis that answers the question you are asking.
- Badges
- Science topic