GMOD / ucsc-hub-js

read and write UCSC track and assembly hub files in node or the browser
MIT License
5 stars 2 forks source link

Connect to encodeproject.org hub #18

Open cmdcolin opened 3 years ago

cmdcolin commented 3 years ago

encode project.org has dynamically generated trackhubs

https://www.encodeproject.org/batch_hub/type%3DExperiment%2C%2Creplicates.library.biosample.donor.organism.scientific_name%3DHomo%2Bsapiens%26biosample_ontology.organ_slims%3Dheart%26assay_title%3DTF%2BChIP-seq%26files.file_type%3DbigWig%26assembly%3Dhg19/hub.txt

It currently throws an error on parent track lacking a shortLabel

So going to here, selecting visualize, automatically can open up a UCSC link

https://www.encodeproject.org/search/?type=Experiment&replicates.library.biosample.donor.organism.scientific_name=Homo+sapiens&biosample_ontology.organ_slims=heart&assay_title=TF+ChIP-seq&files.file_type=bigWig&assembly=hg19

garrettjstevens commented 3 years ago

It looks like the trackDb in this hub has some good examples of how the trackDb spec doesn't always line up with how the UCSC genome browser uses trackDbs. The spec says shortLabel is always required in any track stanza, but loading this up in the UCSC browser shows that shortLabel is not required on composite parent track stanzas.

cmdcolin commented 3 years ago

In autoSql, I experienced similar things where the nicely defined "autoSql grammar" created by the UCSC group is not compliant even by files on ucsc.org. It may be better to try to allow what is in the wild, unless we want to make a plea for the "standard" compliance to https://github.com/ENCODE-DCC