Hello esteemed CellProfiler Team!
I have only recently started working with CPA 2.0 (let me thank you, I wish I had done so earlier), and obviously I have questions
Let’s assume my CP pipeline (and exported SQLite db) contains multiple measurements and images per set (e.g measurements of DNA, proliferation marker, auxiliary nuclear marker). Is it possible to define settings that govern which images or measurements may (or more importantly, may NOT) be used for finding classifier rules? Sometimes I find the “Objectnumber” listed among the rules, at other times I wish to simply ignore a measurement or image (and compare it to my preliminary classification later). Removing these via “Edit Rules” would need to be repeated after every training cycle.
Is there a possibility (or any workaround) to create two independent classifications per database? (e.g. classify cells according to marker A, then create a second classification table and classify them via marker B (and later confront the two independent classifications?). Currently, I process two (or more) markers in separate projects and databases, then classify cells and export .csv tables seperately, finally joining them in Excel for comparison (and after getting used to CPA, I’d really like to avoid spreadsheets =D). Can I link two (or more) per_object classification tables from seperate databases? (Or alternatively, do you think I could load my jury-rigged spreadsheet into the database and use it in CPA?)
The columns of the (compound) classification tables I would like to produce might resemble:
Image_number, Object_number, Marker_A_class_number, Marker_A_class_name, Marker B_class_number, Marker_B_class_name (etc etc).
- Thumbnails (edit - found solution)
I appreciate your tools and your help! Also I can’t help but think there is a simpler method I failed to recognise…