I have added custom build identifiers to the builds.txt file, however after I restart galaxy they are not present in the build dropdown menu. Is there another file I must edit?
Below are the additions
GRCh38 Human Dec. 2013 (GRCh38) (GRCh38)
GRCh37 Human Feb. 2009 (GRCh37) (GRCh37)
hg38 Human Feb. 2009 (GRCh37/hg19) (hg19) - this was a test
To double check, the file being altered is galaxy/tool-data/shared/ucsc/builds.txt.sample.
And this customization is not to create a "Custom Build" for a "Custom Reference Genome", correct? (re: https://wiki.galaxyproject.org/Support#Custom_reference_genome & https://wiki.galaxyproject.org/Learn/CustomGenomes#Custom_Builds)
This line as constructed is problematic. hg38 and hg19 do not match. That will definitely cause at least one conflict in the database, and possibly, reformatting the other duplicated content will resolve all issues.
hg38 Human Feb. 2009 (GRCh37/hg19) (hg19)
If you really want all four in there, try this. It has matching dbkeys in all fields and any duplicated content between the lines removed.
GRCh38 Human Dec. 2013 (GRCh38) (GRCh38)
GRCh37 Human Feb. 2009 (GRCh37) (GRCh37)
hg19 Human Feb. 2009 (hg19) (hg19)
hg38 Human Dec. 2013 (hg38) (hg38)
It could also just be formatted like:
GRCh38 Human Dec. 2013 (GRCh38)
GRCh37 Human Feb. 2009 (GRCh37)
hg19 Human Feb. 2009 (hg19)
hg38 Human Dec. 2013 (hg38)
Or, it can include the source, like:
GRCh38 Human Dec. 2013 (NCBI/GRCh38) (GRCh38)
GRCh37 Human Feb. 2009 (NCBI/GRCh37) (GRCh37)
hg38 Human Dec. 2013 (UCSC/hg38) (hg38)
hg19 Human Feb. 2009 (UCSC/hg19) (hg19)
Let us know how this works. I am testing this in both a 16.07 and the upcoming (but not stable and released yet!) 16.10 default master local. If any problems come up, I'll update.
Make sure that you are also up-to-date. I didn't ask before, but you are running 16.07, correct? The .sample has been removed from the file? Have you made changes to this file before and had it work, or is this the first time? Is there a specific reason why the genomes are being added to the builds list directly, instead of installing the genome with a data manager? This one creates the "dbkey" and populates the genome list in the UI plus creates the .len files needed by many tools (this would be need to be done manually the way you are setting it up now): https://toolshed.g2.bx.psu.edu/view/devteam/data_manager_fetch_genome_dbkeys_all_fasta/86fa71e9b427
If you have more problems, could you post the entire file contents in a format such as a gist? Or a graphic of an open vi session with hidden characters shown. Any method that will deliver all of the content (printed and hidden) exactly like your file.
Apologies for the many questions but let's try to solve this so you can move forward! Jen
Which menu? One or both? The menu in the upload tool or the menu on the Edit Attributes page (found by clicking on a dataset's pencil icon).
The edit attributes page and upload page. The upload page only shows one I modified for hg19 and no GRCH37 or hg38.
It seems the issue is that it is not recognizing hg38 or GRCh37 as a dbkey.
It could be because of content. What happens if you edit the last line to be like:
hg19 Human Feb. 2009 (hg19) (hg19)
I am assuming that you checked for extra tabs, spaces, trailing empty lines. The formatting is very strict. If not, correct that also as needed.
That one does appear. Let me try using that formatting.
hg38 and GRCg37 will still not appear