journal menu
![[Figure 1]](qe5007fig1mag.jpg)
|
|
Figure 1
Venn diagram showing the various clustering methods used to split models into distinct structural units and included in Slice'N'Dice. Shown in teal are clustering methods included in scikit-learn that cluster based on Cα-atom coordinates. Shown in orange are clustering methods included in cctbx that cluster based on the predicted aligned error (PAE) from AlphaFold2. On the left are all of the clustering methods that require the number of clusters to be specified and on the right are clustering methods that automatically determine the number of clusters. The cctbx PAE methods automatically identify clusters: however, if a user defines a maximum number of splits, Slice'N'Dice performs an additional step to merge the closest clusters until the number of splits is less than or equal to the maximum number of splits. |
![[Figure 1]](qe5007fig1.jpg)
 | STRUCTURAL BIOLOGY |
ISSN: 2059-7983
access
