Xcluster: Difference between revisions
No edit summary |
No edit summary |
||
| Line 4: | Line 4: | ||
Xcluster is a program available in maestro, which allows you to cluster the results of a macromodel search. That is, the program analyses the distances in the given geometries and makes groups of geometries. | Xcluster is a program available in maestro, which allows you to cluster the results of a macromodel search. That is, the program analyses the distances in the given geometries and makes groups of geometries. | ||
There are several tools to control the behaviour of the clusters depending on the threshold chosen. The issue is that the program tends to join the species in very big and very small groups. This creates big clusters with lots of geometries (some of them belonging to different isomers), so the program should be used carefully, and using very low thresholds. | There are several graphical tools to control the behaviour of the clusters depending on the threshold chosen. The issue is that the program tends to join the species in very big and very small groups. This creates big clusters with lots of geometries (some of them belonging to different isomers), so the program should be used carefully, and using very low thresholds. | ||
The results of one clusterization can be saved, and use the Xcluster iteratively on different searches with low threshold. It must be stressed that good clusterization is not ensured. | The results of one clusterization can be saved, and use the Xcluster iteratively on different searches with low threshold. It must be stressed that good clusterization is not ensured. | ||
Revision as of 12:30, 26 February 2010
go back to Main Page, Computational Resources, Chemistry & More, Computational Codes, MACROMODEL
Xcluster is a program available in maestro, which allows you to cluster the results of a macromodel search. That is, the program analyses the distances in the given geometries and makes groups of geometries.
There are several graphical tools to control the behaviour of the clusters depending on the threshold chosen. The issue is that the program tends to join the species in very big and very small groups. This creates big clusters with lots of geometries (some of them belonging to different isomers), so the program should be used carefully, and using very low thresholds.
The results of one clusterization can be saved, and use the Xcluster iteratively on different searches with low threshold. It must be stressed that good clusterization is not ensured.