How to set up a new automatic gating analysis by running inference using trained model


The Automatic gating algorithm in the Cytobank platform enables users to train a model that can automatically identify cell populations. You need access to a trained gating model before you can apply it to the new data. This can be achieved by training a model yourself or by another user sharing an experiment with their trained model. Note: the new files can be in the experiment where you trainedthe model or in another experiment. Please ensure that data is appropriately pre-processed with respect to scale adjustments, compensation and panel assignment/channel names. The following instructions show how to apply a trained model to new FCS files.


Step-by-step instructions to run inference using a trained model

1) Go to the experiment withthe new FCS files and then go to the  Advanced Analysis > Automatic gating



2) Choose Run inference using model and then click Create to create one new analysis job


3) Select FCS Files thatyou want to run inference with: In this step, any FCS files that you want to analyze by Automatic gating analysis should be chosen. 


4) Select a trained Automatic gating model: In the earlier training model step, you have created several Automatic gating training models; now you can choose one from the dropdown menu. 


5) (Optional) Enable the checkbox for Clone gates from current experiment: When the box is checked, the Cytobank platform will automatically clone manually defined gates from the current experiment into the child experiment generated by the inference analysis. With those cloned gates, you could make a direct comparison between manual and automatic gating in the illustration page after the Autogate training analysis is completed.


6) Select the appropriate Compensation and Data Scaling settings: 

Please use appropriate compensation and scale settings in the current experiment that contains the new samples.  Please go to the Compensation and Scaling pages in the Cytobank platform to make sure the setting is appropriate. 

7) Select the green Run Autogate inference analysis button at the top of the setup page to start population prediction. You will receive an e-mail notification once the child experiment has been created.


Once inference is completed, you can access the results by clicking the link in the notification e-mail or via the experiment navigation bar and the Experiment Details page.


Please refer to this article about How to interpret Automatic gating results.

Have more questions? Submit a request