genome browser SNP-displaying bug

[Vadim Mihailov]

Hello. 

Just found something in the genome browser that may be a bug and is pretty disturbing.

1. I've made a search using blat and got several links with the same score - becasue this were highly similar genes. I checked what is that I needed and proceeded to a browser window.

2. My goal was to check is there a SNP at the end of my sequence or not. I keep all SNP tracks toggled on but compressed to a "squish" state.

3. I saw a SNP at the target point. To proceed, I changed this SNP track to a "full" state and the SNP disappeared.

4. Some time later I found that this was SNP track from another gene. My target gene is KRT14, and after expanding its SNP track (actually, all of them) refreshed as tarcks from KRT16 (a highly similar gene). The links from these SNPs to dbSNP actually lead to KRT16. This can be observed, but is very easy to miss if you don't expect such thing.

5. After I made blat search again and tracks were already in expanded states (as full) they represent correct tracks from an actual gene.

Just a check if it is known to your team. 

The sequence I used:

GGGGGGCGGCTATGGCGGTGGCTTCAGCAGCAGCAGC

--

Regards, Vadim Mihailov

[Chris Villarreal]

Hello Vadim Mihailov,

Thank you for informing us about this issue. Unfortunately, we are unable to reproduce the problem locally. Can you please provide us with a link to a saved session that shows the problem? Thanks!
If you have any further questions, please reply to gen...@soe.ucsc.edu. All messages sent to that address are archived on a publicly-accessible forum. If your question includes sensitive data, you may send it instead to genom...@soe.ucsc.edu.

-Chris V
UCSC Genome Bioinformatics Group

--

---
You received this message because you are subscribed to the Google Groups "UCSC Genome Browser discussion list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to genome+un...@soe.ucsc.edu.

[Chris Villarreal]

Hello Vadim Mihailov,

Thank you for providing a video of the error. The culprit here seems to be the browser cart (hgsid). When you open the links from BLAT by right-clicking and selecting "open in new tab", each Genome Browser tab will share the same session id, which causes any changes in one tab to overwrite changes in the other. You can verify this by only opening the first link from your BLAT results and changing the display from squish to pack, where you will see the correct SNPs, whereas if you have both tabs open then you will see the situation you describe. If you would like to have two separate tabs, then after opening a second tab from the BLAT results, change the URL to get rid of everything except forhttp://genome.ucsc.edu/cgi-bin/hgTracks , then hit enter, this will assign a new hgsid for that tab and the session should act independently.

If you have any further questions, please reply to gen...@soe.ucsc.edu. All messages sent to that address are archived on a publicly-accessible forum. If your question includes sensitive data, you may send it instead to genom...@soe.ucsc.edu.

-Chris V
UCSC Genome Bioinformatics Group

On Sun, Nov 20, 2016 at 1:03 AM, Vadim Mihailov <mihailo...@gmail.com> wrote:

Recorded a video where I compare SNP tracks for KRT16 and KRT14. Different SNP IDs are shown, After expanding, ID changes, After refreshing, it changes back to normal. Both genes are opened in the same web browser and both pages are present during this manipulation.

SNP tracks must be compressed before the search.​

 SNP tracks display bug.avi

​

On Sun, Nov 20, 2016 at 11:51 AM, Vadim Mihailov <mihailo...@gmail.com> wrote:

Checked it, found again.

Maybe both genes have to be opened in the browser. I'll try to make some visual reference.

--

С уважением, Вадим Михайлов

--

С уважением, Вадим Михайлов