Hello DepMap team,
I am seeing a lot of cell lines with copy number values of 0 since the 2020 Q4 releasing.
A lot of these cell lines did not have these low copy number values a few releases ago.
Can you check if this could be a mistake?
Hello DepMap team,
That certainly sounds suspicious. We’ll investigate and will let you know once we understand whether this is accurate or a mistake.
I was wondering if you have any updates on this issue.
Sorry for the delay! Are you seeing genes for which the copy number changed dramatically for most cell lines, to a lower level?
I will also say that there is a couple things to know with this new release that might have an influence in the number of zeroes.
We used to only report CN on genes where we had reads. Now that we have WGS, we have reads for genes that are not reported by ICE or AGILENT (our WES targets). So that makes for some genes being 0s in most samples. Also we have added Y chromosomes… so lots of 0s as well for XX samples.
Finally we have decided to not release this new reprocessing for all samples where the correlation to the previous release was below 0.85 -to make sure that there is nothing weird for these samples-. Using the previous version meant that a couple of genes that were not reported in previous quarters, the value is now NA.
In any case, if changes create any issues for you, you can use the previous release, which are still available.
To add to this thread, I found a relevant example.
Some well-known ERBB2 amplified breast cancer cell lines have a CN count of 0 in the 20Q4 release (e.g. BT-474 and HCC1954). In previous releases these cell lines had a high CN count (5.78 for HCC1954 and 3.79 for BT-474 in the 20Q3 release).
Hope this helps in solving the issue.
Thanks all for reporting this issue. Please see the following post for more information:
I was using the 20Q1 data for an analysis in my project. Can you please let me know if 20Q1 version does not contain this bug?
@Arkajyoti this issue only affected 20Q4, so 20Q1 is Ok to use.