Add S. Cerevisiae June 2008 genome to IGB Quickload - User request

Create and test Quickload containing just these data and notify Dr. Loraine when it is ready to deploy.
Ensure appropriate synonyms are added to synonyms.txt and species.txt bundled with IGB.

I used the files attached here.
I changed the name of the .2bit file to match the folder. The name of the bed file can stay as-is.

In order to create an index for the bed file I had to unzip it and sort it.
sort -k 1,1 -k2,2n S_Cerevisiae_Jun2008_sgdgenes.bed > S_Cerevisiae_Jun2008_sgdgenes.sorted.bed

I made the genome.txt file as suggested in the users guide:
twoBitInfo S_cerevisiae_Jun_2008.2bit genome.txt

Both the genome.txt file and the set of chromosomes in the bed file are as follows:
$ gunzip -c S_Cerevisiae_Jun2008_sgdgenes.bed.gz | cut -f 1 | uniq | sort
$ cat genome.txt | cut -f 1 | sort

2micron
chrI
chrII
chrIII
chrIV
chrIX
chrM
chrV
chrVI
chrVII
chrVIII
chrX
chrXI
chrXII
chrXIII
chrXIV
chrXV
chrXVI

I don't know what the 2micron sequence is supposed to be.
Since the annotations and the genome use the same chromosome names, I don't need to add a chromosomes.txt file.

synonyms.txt

To find other terms that might be used for this genome version, I looked at the UCSC ftp downloads site.

In this document, I found that the following terms for the 2008 genome: sacCer2, SGD June 2008
ftp://hgdownload.soe.ucsc.edu/goldenPath/sacCer2/chromosomes/README.txt

In this document, I found that the 2011 version (already supported in IGB) should have these synonyms: SacCer_Apr2011, sacCer3
ftp://hgdownload.soe.ucsc.edu/goldenPath/sacCer3/chromosomes/README.txt

One the most likely times for needing a genome synonyms is when loading data from Galaxy.
I went to useGalaxy.org, and selected NGS: RNA Analysis > TopHat
I selected to Use a built-in genome, and entered "Cerevisiae" to search within the pulldown menu for genomes. Options are:
Yeast (Saccharomyces cerevisiae): sacCer2
Yeast (Saccharomyces cerevisiae): sacCer3

So I think I have all the synonyms I need.
_________________________

species.txt

I don't when we need this. Its in the instructions for creating a quickload site, so I made it. I supplied sacCer as an alternative genome prefix.

HEADER.md

I stole the HEADER.md file from the S_cerevisiae_Apr_2011 directory in the SVN repo.
Modified it to reflect the current genome version and data sources.

_______________

annots.xml

I stole the annots.xml from from the S_cerevisiae_Apr_2011 directory in the igbquickload site.
Modified it to use the bed file name for this genome.

I set up an AWS EC2 instance and followed the instructions on this page to set up Apache.
https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/CHAP_Tutorials.WebServerDB.CreateWebServer.html
Section: Install an Apache Web Server with PHP
Followed these steps, except in step 3, I just used "sudo yum install -y httpd" and ignored the rest.

You can now access the quicklaod site for the 2008 S.Cerevisiae genome assembly here:
http://18.222.191.240/Quickload_IGBF-1189_S.Cerevisiae/

I tested this as a data provider in IGB and it looks great.
I have not tested the species.txt or synonyms.txt file.
The only thing that's not working is the HEADER.md file. Apache seems to be completely ignoring it, doesn't even list it as a file in the directory.

I opened the HEADER.md file in RStudio and made a HEADER.html file from it, uploaded that to the EC2 instance.
Now when go to that location via my web browser I see the html page. But the overall style is very different from the main IGB quicklaod.

I think this is enough for Ann to be able to test the files. There's no reason to spend time adjusting my web server settings, the purpose of this site is to test the quickload files. It might be nice to know what the right settings are (for general reference), but there's no reason to hold up this issue.
Ann, for future reference, how could I set up my server to have the general style/appearance of the main IGBquickload site?
_______________________

I am assigning this to Ann to review.

You can now access my quicklaod site for the 2008 S.Cerevisiae genome assembly here:
http://18.222.191.240/Quickload_IGBF-1189_S.Cerevisiae/

If no corrections are needed, then I think this is the point where I just hand it off to Ann completely.

DO let me know when you are done using the test site. I will shut down the EC2 instance.

Add URL attribute to annots.xml; use relative link as in human genome annots.xml

Gene models do not appear to have functional information. Please add description and gene name info into fields 13 and 14, similar to what we do for the human genome. The descriptive text for each gene should be available from one of the Entrez gene files. Search the IGB wiki for documentation describing how and where to get functional descriptions for a genome. The documentation is now a few years old and so some of the file names and locations may have changed. However, the basic idea should still hold, which is: we use the Entrez gene databases from NCBI to obtain succinct descriptions of genes and show them to users when they select gene models.

added url.

I'm trying to find a file with the gene descriptions.

This file:
https://downloads.yeastgenome.org/curation/chromosomal_feature/saccharomyces_cerevisiae.gff
has gene symbols (which do not match our bed file), and aliases (which do seem to match), and descriptions. The locations given in this file are based on genome (version=R64-2-1), which is SacCer3. Since we have a bed file for SacCer2, and the descriptions should be the same if the name is the same, we should be ok to use this. It would be nice to find annotations that were designed for the SacCer2 assembly, or a file that has matching names as the primary name (not buried in a list of aliases).

The gene descriptions were added following the instructions here:
https://wiki.transvar.org/display/igbdevelopers/Updating+RefGene+UCSC+data+set+for+an+existing+genome+in+IGB+QuickLoad
With some modification.

I encountered a problem with the ucscToBedDetail.py script from the GenomeSource repo:
https://bitbucket.org/lorainelab/genomesource
That issue has been resolved.

The authoritative body for this genome is SDG. UCSC provides a bed table of sdg genes. The identifiers used in this bed file are locus_ids, which is an available field in the gene_info file from ncbi. I modified the ucscToBedDetail.py script to filter by taxID and use the locus id rather than the gene id to get the information from the gene_info table. I copied the modified script (ucscToBedDetail_SacCer_locusID.py) to the test quickload site so Ann and include it as documentation if appropriate.

Out of 6,717 models in the sdg genes bed file, 698 were not found in the gene_info list (so they have NA for the gene description).

This is ready for review.

Remember, you can access the quicklaod site for the 2008 S.Cerevisiae genome assembly here:
http://18.222.191.240/Quickload_IGBF-1189_S.Cerevisiae/

I am assigning this to Ann for review.
Please see my last comment.
If this passes review then Ann can take over.

• Added to new https://svn.bioviz.org/repos/genomes subversion repository. To browse, visit https://svn.bioviz.org/viewvc (visit on non-UNCC network due to issues with internal DNS, to be fixed soon)
• However, did not add S_cerevisiae_Jun_2008_refGene.bed.gz because it is not mentioned in the annots.xml file.

Deployed as part of current (newly migrated) quickload site.

To test:

• Test using master branch installer for 9.0.2
• Inspect Species menu of Current Genome tab; check for new yeast genome related menu options (IGB has a bug whereby newly added genomes are shown incorrectly using their UCSC-style or other synonyms)
• Select Current Genome tab, select Saccharomyces cerevisiae species (this is budding yeast)
• Check that 2008 genome is listed in genome version menu
• Select 2008 genome
• Make sure that reference gene model data set is loaded upon visiting the genome
• Make sure that sequence data can be loaded
• Make sure that the menu item in Current Genome lists the species and genome correctly, with correct tooltips

Report any problems here and if there are problems, return to To-Do column.

Tested as described, the application behaves as expected.