have two questions:
is there a way of exporting the dendrogram of the MCS library clustering to
excel (for structure visualisation of every molecule)?
is there a way of selecting a centromere (most representative
molecule) form each cluster?
help will be much appreciated!
The answers for your questions:
1. The only way you can export the cluster data to Excel is to export it into an SD file in libmcs. This SD file will contain all clustering data, which you can put in JChem For Excel.
2. You can't select the centroid in libmcs, because the representation does not allow that. You should use Sphere Exclusion clustering instead. It that case centroid can be easily selected.
I hope this helps.
Thank you very much for the reply. What is this Sphere exclusion clustering? Is it in JChem? Could you please give me some more info cause I am new to the software?
I have finished with clustering in libmcs. I now want to find centroids from the large clusters. As I think I about it, I have to import the molecule of each clusters and find centroid. Is there any guide for this please?
Thank you very much for your reply!
It took me a while to reply because of heavy schedule.
I confused the centroid with the 'most representative molecule ' of a cluster. So what I need to find is the most representative molecule for each of my clusters. I discovered that you can a get a similarity/dissimilarity matrix for the molecules of a cluster by using the command line.
I am using this command: jcsearch -q [file name] -f [output file name] -t:i:1
but I get an error message that is not recognizing .smiles or .sdf files. What kind of file extension I need to use please?
Any advises on how to finish this task please?
I want to obtain a similarity matrix between a set of molecules. I am using the following command in cmd
screenmd file1.sdf file2.sdf -t:i:1 -o output.txt
but it doesn't work, could you please let me know what I am doing wrong?
For clustering I highly recommend that you use our command line tool jklustor. You can use the Sphere Exclusion clustering algorithm with that, which will hopefully suit your needs.
To get the full manual of jklustor under the command line, type
I also link here the official JKlustor manual:
Answering your question regarding the screenmd, you used a wrong option with the command. Please see my answers regarding that in my previous answers for you.
Thank you very much!