CHESCA Covariance Analysis

CHESCA Covariance Analysis

A minimum of four states is recommended for CHESCA and at least two replica HSQC spectra are required for each state. Among all the states for CHESCA, it is advisable to ensure the presence of both active, intermediate and inactive states. To input each state for CHESCA analysis, designate a state name and select all the corresponding spectra in the spectrum list. Click the “Add” button to add the state and repeat to add all the states for the analysis. For the SVD analysis, an antagonistic state is recommended as the reference state. Select the chosen reference state in the state list and use provided arrow to move the reference state to the top. By clicking the “Save” button, you can save this state list for later use.

Similar to CHESPA, a correction factor for the N dimension is required for the ¹⁵N ppm values. We recommend the ratio of 0.2. The correlation coefficient for the covariance matrix and the agglomerative clustering is defaulted at 0.98. The N and H dimension cutoffs are set at 5 and 10 Hz, respectively, to exclude possibly insignificant chemical shift differences. We recommend these default values at least for preliminary CHESCA analyses. An option similar to the inter-peak filtering scheme of CHESPA is also available for CHESCA. Once the parameters are set, click “Run” to compute the correlation matrix (R). When the job is finished, an output folder of raw data will appear in the subfolder “CHESCA” in the Sparky folder. You can rename the folder according to preferred naming conventions. Once the job is finished, the system will prompt to plot the correlation matrix. In addition, you can also view the R matrix by loading the data folder by clicking the “Browse” button in the CHESCA-SPARKY importer and clicking the “Plot R” button. The pair-wise correlations will also appear under the CHESCA tab. Double click any pair wise correlation if you wish to view the linear correlation of the states between a selected residue pair.

To start the agglomerative clustering and SVD analyses, simply load the data folder through the CHESCA-SPARKY importer and click the “Check” button. This will initiate the agglomerative clustering in the single linkage (SL) and complete linkage (CL) modes as well as the SVD analysis. The SVD analysis is implemented in two modes, the referencing mode and mean centering mode. The referencing mode will use the first state in the state list as a reference. You can view the clustering and SVD results under the tabs of CHESCA-SL, CHESCA-CL and SVD by double clicking the clusters or principal component plots. A PyMOL function is integrated in the CHESCA-SPARKY to review the agglomerative clustering results. You can choose a pdb file of your choice after clicking the “PyMOL” button. CHESCA-SPARKY will create a pdb file in your data folder called “chesca.pdb” which will realign the input pdb file with the assignments from the spectra. A PyMOL command line is copied to the clipboard as well. If the PyMOL program does not start automatically, you can simply start PyMOL and paste the command line. To show individual clusters, select the clusters you want to see in the result window and type “@p” in the PyMOL command-line to update the structure.

Video Tutorials

<< Previous: CHESPA Projection Analysis

<<< Back to Home