Astrolabe is using the Ek'Balam clustering-and-labeling algorithm in order to assign cells to canonical populations. We refer to this classification as the "Assignment" level. Astrolabe has a default labeling hierarchy that was curated from Maecker et al., Nat Rev Immunol. 2012 Feb 17;12(3):191-200 and Finak et al., Sci Rep. 2016 Feb 10;6:20686.
It is possible that the hierarchy is not the best fit for your panel. For example, maybe you want to use a different marker to identify a particular population or maybe Astrolabe did not identify a population of interest (macrophages are a common example). In such cases, we can customize the hierarchy for you. Contact us at firstname.lastname@example.org for details.
The "Assignment" level is designed to capture what is known and established about your panel. As a rule of thumb, you should consider customizing the hierarchy if one of the following two conditions are meet:
- You designed the panel with a particular population in mind whose immunophenotype follows a well-defined set of markers.
- There is published literature showing that a well-defined set of markers can identify one or more populations in your data.
If your population-of-interest is nebulous, has not been characterized in the past, or requires novel discovery, we strongly recommend not customizing the hierarchy. Instead, rely on the "Profiling" level or the Differential Expression Analysis. Email email@example.com if you have further questions. You're also welcome to review the following video: