wired bug: not showing hammock track after changing name

[biomystery]

The below is using the washU epigenome browser v46.1

1st link/json is OK to show hammock track: 1st link/json

2nd link/json failed to show hammock track:2nd link/json with error msg:

Parsing URL parameters... Your host: epgg4 Server load: 0.06, 0.03, 0.05 hg38 genome loaded Setting view range... Parsing hub file... Loading 3 tracks... Tracks added hg38 Missing track for reordering: 10122371531454699 Stand by

The difference between the json of 1st and 2nd link is just the name of hammock file:

$diff ok.json failed.json                                                                                                                                          
 8c8                                                                                                                                                                                                         
 <    "url": "http://epigenomics.sdsc.edu/zhc268/test/ok.narrowPeak.hammock.gz",                                                                                                                    ---                                                                                                                                                                                                         
>    "url": "http://epigenomics.sdsc.edu/zhc268/test/failed.narrowPeak.hammock.gz",

The ok.narrowPeak.hammock.gz and failed.narrowPeak.hammock.gz are exactly the same ( by cp -p ok.xxx failed.xxx .

zdiff ok.narrowPeak.hammock.gz failed.narrowPeak.hammock.gz
diff ok.narrowPeak.hammock.gz failed.narrowPeak.hammock.gz
diff <(zcat ok.narrowPeak.hammock.gz) <(zcat failed.narrowPeak.hammock.gz )

===== This is really puzzle me.

Because start from 4/26/18, the hammock produced by my pipeline cann ot be shown. Before that date, the hammock files generated by my pipeline were ok to be seen. I started to dig why and found this phenomena. Thanks!

[lidaof]

Hi @biomystery Frank, seems the index file .tbi is not existed with the .gz file, can you please put them back and try again? Thanks for using the browser.

[biomystery]

added tbi by tabix -f -p bed. still not working. previously the "ok" was showing but after I changed a folder, both not working

$tabix --version
tabix (htslib) 1.7
Copyright (C) 2018 Genome Research Ltd.

[lidaof]

I can see your tracks by doing refresh cache, if the .tbi file changed, we need a manually refresh cache, there is an App called validate datahus for this purpose. You should be able to see both your hubs now.

[biomystery]

Hi Daofeng,

Thank you for following up! I found ok.json works but failed.json still not show hammock track. I am not clear about what changed and what should I do at my side.

1. refresh cache: do you mean clear my browser cache and reload?
2. validate datahubs: do you mean manually refresh at the server side?

So in future, how to avoid cache issue if I want to use WashU browser's datahub function? Or should I install the browser at my server and then can do the server-side operations to avoid this?

Thank you in advance for your advice. Best Frank

[lidaof]

sorry for the confusion Frank. It's the function in the browser I meant, for a custom track, usually you can right click and select Refresh cache:

if you have many tracks in datahub, you can do that in batch from the Apps menu:

this happens usually when a track file updated later but the browser cached old index file, currently we need manually refresh.

[biomystery]

Thank you, Daofeng. You've answered all my question!

Best Frank