journal menu
![[Figure 9]](qe5007fig9mag.jpg)
|
|
Figure 9
Pipeline following Slice'N'Dice, made possible by the CCP-EM software suite (Burnley et al., 2017  ) GUI Doppio (Burnley et al., 2023). In the suite, Slice'N'Dice is split into three jobs: Slice, Dice and Slice'N'Dice. The model is a ColabFold model of the SARS-CoV-2 spike protein PDB entry 7ymt generated using ColabFold (Mirdita et al., 2022) with pLDDT scores stored in the B factor column of the PDB file. Residues under a certain threshold (70; the default and the value used in generation of the figure) are trimmed and the remaining residues undergo the slicing process. In this figure, two separate slice jobs were performed using BIRCH and k-means clustering (Pedregosa et al., 2011). (n.b. the colours for each slice are inconsistent between jobs and split numbers. A colour key is supplied with each run to indicate which filename links to each colour.) For the purpose of this exercise, k-means split 4 was used without slice 1 as this slice contains extraneous features. The Dice job is run next using the input models from the Slice job. An input map is used as the template for the map fitting. In this example, this is a 6.55 Å resolution reconstruction of the spike protein from the EMDB (EMD-33942). After the map fitting, two output windows are displayed in Doppio, one containing the models that passed the classifier check and a second output of the combined `passed' models and any remaining models. These are coloured according to the confidence score of the classifier, an extra output for the Dice job. |
![[Figure 9]](qe5007fig9.jpg)
 | STRUCTURAL BIOLOGY |
ISSN: 2059-7983
access
