Large, single gene queries crash on Chrome, but not Firefox

[Zachary Whitfield (WHITFIEZ)]

Hi,

Some colleagues and I have been troubleshooting crashes we see on our own instance of cBioPortal.

When using Chrome to query a highly mutated gene on a fairly large dataset (~28,000 samples corresponding to ~35,000 alterations), Chrome will hang at the "Annotating with OncoKB and Cancer Hotspots" for ~3-4 minutes. It will then just freeze, or display a mostly blank page (still has some of the header images and tabs, but no actual oncoprint). Monitoring the network requests within the browser shows one of the 'fetch' steps as "(failed)" 

However, Firefox consistently returns results for the same query.

Performing a query with a less mutated gene (~28,000 samples corresponding to ~23,000 alterations), returns results in both browsers no problem.

Has anyone experienced a similar issue? Are there any ideas why Chrome might fail here? Please let me know if I can provide any more information that would be helpful (or if it would be better to post this as an issue on GitHub).

Thanks so much,

Zach

[JJ Gao]

Hi Zack,

Thanks for contacting us. Was this problem on the public portal? If so, could you send us the URL after submitting the query? 

-JJ

--
You received this message because you are subscribed to the Google Groups "cBioPortal for Cancer Genomics Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cbioportal+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cbioportal/ca20d837-c585-4aa9-a611-a3ab0f3ba775n%40googlegroups.com.

[Zachary Whitfield (WHITFIEZ)]

Thank you so much for responding JJ. This was not on the public portal, but rather our own deployed instance with our own internal datasets (I realize that makes it more difficult to help troubleshoot).

It seems we have this issue when the query will return a large (>~30,000) number of alterations for a single gene. I was unable to find a single gene to return such a large number of alterations on the public site (though I definitely could have overlooked something). I was, however, able to return that large of a query on the public Genie site with no issue on Chrome.

Any information or clues on why we might be experiencing such an issue would be greatly appreciated. Are you aware of any difference with how the public site (or an internal deployment based on the public site) fetches data (in particular the OncoKB annotations) compared to the public Genie site?