I have lists of cell lines loaded in my account and when I now do a ‘Custom Analysis’ and pick a sublist the sublist is ignored and all cell lines are analyzed anyway. If I open the results in Data Explorer 1 or 2 I have the same output the only difference is that in Data Explorer 2 the cell lines are automatically restricted to my list on the dot plot but the underlying results are that of the analysis of all cell lines.
Thanks for reporting you’re seeing a discrepancy. I’d like to understand what’s going on here.
I’ve tried a few things but I’ve been unable to reproduce the behavior I think you’re describing. However, I haven’t been able to find a difference between the two:
They seem to be returning the same result. When you saying " the underlying results are that of the analysis of all cell lines." are you saying the volcano plot looks different for some analysis you ran between DE2 and DE1?
Thanks,
Phil
I will walk through what I see.
1a) This is an analysis done on all cell lines.
1b) The output in Data Explorer 1 with table view and the table sorted with strongest positive correlations first.
1c) The output in Data Explorer 2 with table view and the table sorted with strongest positive correlations first.
2a) This is an analysis done on a subset of Glioblastoma cell lines that were freshly defined as a model using DepMap IDs.
2b) This is the Data Explorer 1 default output of this analysis with the table viewed and sorted with strongest positive correlations first. Notice this table output is identical to that of the all cell line analysis.
2c) The same Data Explorer 1 output with my Glioblastoma subset now labeled.
2d) This is the Data Explorer 2 output with the Table again shown and sorted as before. Notice this table output is still identical to the all cell lines analysis. By default Data Explorer 2 filters the dot plot to the Glioblastoma subset cell lines although the analysis seems to have been conducted on all cell lines.
2e) The Data Explorer 2 view of the dot plot with the Glioblastoma subset filter turned off.
So as you can see this particular analysis is not being restricted to the Glioblastoma subset when the parameters for the analysis are initially set. I have identical results in both Firefox and Chrome.
Thank you for the detailed explanation.
I have successfully reproduced the issue and have a good lead for tracking down what is going wrong, and so far it points to our changes to the storage backend as being the most likely cause.
I’ll update this thread once I have more information and/or a fix for this.
Thanks,
Phil
Thanks Phil! Look forward to this working again.
Hello,
A fix for this problem has been deployed this morning.
The erronous behavior manifest when one of the following datasets were selected due a bug in our new storage backend:
Batch corrected Expression Public 24Q2
Copy Number Public 24Q2
Protein Expression
Sanger Combinations (AUC)
Sanger Combinations (Library AUC)
Sanger Combinations (Library viability)
Sanger Combinations (viability)
Methylation (1kb upstream TSS)
Protein Array
Trying out your example, which referenced Batch corrected Expression Public 24Q2 . I’ve confirmed that the table section on the left reports the expected number of lines as opposed to the 1060 (which is to say, all lines) and appears to now correctly be running on the selected subset.
Thanks again for bringing this to our attention.
Thanks,
Phil
Hi Phil,
Thanks for fixing so quickly; much appreciated!
