Hi
I have simulated a protein for 100ns and the cluster analysis is performed. The xpm file is converted to eps. How should I read a cluster output and what does it infer.
I have performed this analysis to understand the stability of the protein structure.
The x and y axis are not clear.
It is a simple RMSD matrix data, illustrating the conformational deviations of the structure over the MD simulations.
Yea. What does it actually mean? How to interpret it.? Now In the image, the lower diagonal is Red and the upper diagonal is in mixed colour. From this how can I say the formation of clusters and it’s stability? The log file says there are totally 88 clusters. what Does it mean literally? How to find the stable structure from this cluster algorithm?
Sorry for so many questions. This analysis actually confuses me a lot.
I don't think you can comment anything about the stability of the protein based on clustering. Clustering, as the name suggests it just groups similar structures/conformations you sampled through your simulation (most likely based on the RMSD). As you have mentioned that your log file reads "totally 88 clusters", it means that all the conformations you sampled through your simulations can be grouped into 88 clusters/groups and that you can select one representative structure per cluster and perform additional analyses on those structures only to observe whatever your research question is, for example, you can study the difference between these structures and see which region of your structure is highly flexible or varying and then you can carry on with whatever questions you have regarding your target.
Hi Mr. Sanjay
Thank you. That was helpful. Still more clarification please.
I am actually confused to read the log file and the two dimensional plot. Can you help me that?
I think the log file is self explanatory, isn't it?
Anyhow, the log file basically explains your chosen cutoff for the analysis and then it tells you the number of clusters it obtained based on the RMSD cutoff (and also some more information), then it also writes the representative structures (middle structures, for different algorithms refer the corresponding papers) from each cluster to an output pdb file. The table in the log file gives you the time frame of each structure in each cluster (cluster members column) and also tells you the RMSD and the time frame of the middle structure (middle rmsd colum) that it probably wrote to the output pdb file. Hope this helps.
Next, in terms of the xpm file, if i am not wrong, the upper half explains the pairwise RMSD values and the lower half explains whether two structures are in the same cluster or not (most likely some of your clusters are really big i.e. it contains large number of conformations you sampled). If you want to plot this properly, either you have to write your own parser and then plot it (i don't know whether any such ready made script exists, i wrote one such script for parsing and plotting the secondary structure output of gromacs, but it cannot be used directly to plot this) else you can adjust the parameters of the m2p file (play around) and use it as an input in the xpm2ps program (-di option), also you can change the time unit to nano seconds "-tu ns" during your gmx cluster execution.
Here you can find my secondary structure output parser, which you might use it as a starting point to write your own parser!
https://github.com/sanjaysrikakulam/gromacs_utility_scripts/blob/master/gromacs_scripts/convert_ss_xpm_to_png.py
- Badges
- Science topic