Prediction Database

This feature will allow the user to build databases with experimental assigned spectra which can be used to enhance the prediction results of similar structures.

You can build 1H, 13C and X nuclei  (11B, 15N, 17O, 19F, 29Si, 31P) NMR databases from any Mnova document containing a molecular structure with the resulting assignments.

Building the database is really easy, just open your NMR dataset with the molecular structure; make the assignments and follow the menu 'Prediction/Update 1H Prediction DB' (you can also update the prediction database from the Well Plates view of the DB plugin by right clicking on any of the plates, or even directly from a predicted spectrum).

This will display a new dialog box, with information about the molecular structure and the assignments. From here, you can select the DB Letter to save your assignments (from 'N' to 'Z'; the database letter chosen represents the unit which can be excluded from any subsequent prediction or search. Thus, the database letter should be carefully selected on an organization wide basis).

Here we can see the effect of having updated the prediction database. The blue spectrum corresponds with the original dataset, the green one with the original prediction (prior to the update of the Database) and the red one corresponds with the resulting predicted spectrum after having updated the database with the assigned experimental dataset. As you can see, now the prediction is even more accurate!

When running a 1H or a 13C prediction you can drill down into any of the 13C user database predictions (if you do not have licenses of NMRPredict Server, you will only see your own prediction Databases; from N to Z). The details can be accessed through Modgraph Predictions table (by following the menu 'View/Tables/Modgraph NMR Predict/Modgraph Drill down Table') or by clicking on the 'Drill down Table' button from the 'Prediction' ribbon.

After having double clicked on a row, the whole list of shifts used by Modgraph predictor to generate the final predicted shift will be available.

Then, you can access the details of each compound from the database, by clicking on any row in the listed shifts:

Once you have built your assignments database, load a molecular structure with assignments and select any structure search option under the 'Prediction' menu to search for structures and substructures in the DB:

You can also Search for Tags:

Selecting 'Prediction DB Info' will show you the number of records for each nucleus in each prediction DB:

Min and Max values correspond to the number of entries in each database, where the deleted entries are also counted.

Follow the menu 'Prediction/Browse Prediction DB' or 'View/Panels/Prediction DB Browser' to navigate through the database:

From this browser, you will be able to navigate through the different compounds in the DB by using the 'blue arrows'. You can paste the assigned molecular structure into Mnova by clicking on this button

You can edit any assignment by double clicking on the cell and typing the new value. Once you have changed the value, click on this icon to update the changes in the DB. Click on this button to remove any molecule from the DB.

Select another DB letter (from 'N' to 'Z') by clicking on the 'Browse DB' button:

From the Prediction DB Browser you can also run substructures searches, by clicking on the resulting icon for each nuclei:

The location of the prediction database can be selected from the 'Prediction Nuclei Options' menu, after having clicked on the 'Predictor Properties' button.

Please make sure that you have rights to write in the folders above (if not, you will need to change the paths):

Prediction DBs are stored by default in the following locations (with writable permissions).

- 1H: C:\Users\Username\AppData\Roaming\Mestrelab Research S.L.\MestReNova\NMRPredict\NMR1HDatabase

- 13C and X-nuclei: C:\Users\Username\AppData\Roaming\Mestrelab Research S.L.\MestReNova\NMRPredict\csearchdb2nz

However, if you are getting a red path, please try to find the “13C and X-nuclei folder”  called “csearchdb2nz”. Then copy the “csearchdb2nz” folder to any other location where you have writing privileges (for example to your ‘My documents’ folder or directly to your ‘desktop’). Finally select the path in the ‘Prediction Databases Setup’ dialog box and try to update the prediction.

If you have loaded a molecule in Mnova (or a SDF containing several molecules) with assignments, you can update the prediction Database by following the menu 'Tools//NMR Tools/Batch Update Prediction DB':

Follow the menu 'Tools/NMR Tools/Batch Update Prediction DB' to export the assignments that you have in your Mnova Database into the Prediction Database:

This script will be useful for example to import external prediction Databases into a Mnova Database and from here export the assignments to the Mnova prediction DB (in combination with other scripts as it is described in this chapter)