Documentation for workflow and scripts
- 1 Location of directories
- 2 Scripts
- 3 Moving files to Textpresso-dev for production
- 4 Changes to software, files, workflow
- 5 Finding gene popularity (number of mentions in papers)
- 6 Software configuration management
Location of directories
- Scripts are at:
- All scripts write the data to the files here:
above soft-links to: /media/data1/jdone/concise_descriptions/www/concise_descriptions
/var/www/html/concise_descriptions points to the same directory
- GitHub ready files, the latest version (not in development) of scripts:
above soft-links to:
Data at: /data2/srv/textpresso-dev.caltech.edu/www/docroot/concise_descriptions
Scripts at: /data2/srv/textpresso-dev.caltech.edu/www/cgi-bin/concise_descriptions
The software for automated concise descriptions is available online: https://github.com/WormBase/automated_descriptions
The requirements are outlined here: http://wiki.wormbase.org/index.php/Generation_of_automated_descriptions
- Make sure that the files in the data directory , report the current production release, release of sources and the list of Wormbase supported species. These files need to be updated every release:
- production_release.txt holds the production release for the output of the concise descriptions. (needs updating every release)
- release.txt holds the release information for the sources. (needs updating every release), = "production_release" minus 1
- species.txt lists the tab separated values of the species abbreviation, project name, full name and gene prefix for each species (needs updating only when new species are added)
c_briggsae PRJNA10731 Caenorhabditis briggsae Cbr
- Go to the directory that includes the scripts:
- The following files must be edited when there are changes, eg., changing of the IP address of Postgres, change of ace server port, etc (will not happen every release, but watch)
- db_ip.txt is the IP address of the SQL database that holds much of the daily updated information for Wormbase.
- html.txt is the location of the docroot or /var/www directory for the output.
- parallel_path.txt is the path to where GNU Parallel is installed, different for Ubuntu and RedHat.
- cgi.txt is the location of the scripts.
- acedb_port.txt is the location of the port of the acedb server.
- acedb_host.txt is the IP address/hostname of the acedb server.
- In: /home/ranjana/concise_descriptions_scripts
- Create the directories needed by the scripts for a given production release:
$ ./create_release_directories_parallel.pl WS253
(Sometimes 'parallel' makes a mistake, run twice!)
Scripts for source files
- There are input files that must be downloaded and pre-processed, formatted for input; the following scripts must be run:
./biomart_query.pl ./download_gene_lists_elegans.pl ./download_geneIDs_all_parallel.pl ./download_orthologs_all_parallel.pl ./download_gene_associations_parallel_all.pl ./go_terms_only.pl ./get_alt_id_terms_only.pl ./go_obo_to_go_ace.pl ./list_dead_genes.pl ./create_curated_gene_list.pl (queries Postgres) ./list_uncurated_genes.pl ./create_gene_list.pl (queries Postgres) ./parse_gene_lists_elegans.pl ./parse_orthologs_all_parallel.pl ./acedb_gene_class.pl (queries ace server) ./publication_gene.pl (queries Textpresso-dev server, looking at indices, takes 1-2 hours) ./download_wbbt_elegans.pl (downloads the anatomy obo file from the WB FTP site) ./download_anatomy_elegans.pl (downloads the anatomy association file from the ..) ./generate_wbbt_obo_terms_elegans.pl (creates a txt file from the anatomy obo file) ./generate_wbbt_obo_synonyms_elegans.pl (creates a txt file with synonyms) ./00_wbbt_obo_to_wbbt_ace.pl (creates a .ace file with all parents and child terms) ./01_wbbt_obo_to_wbbt_ace_names.pl (same as above with names) ./neuron_xlsx_read.pl (reads the excel spreadsheet from O. Hobert) ./download_hgnc_family_elegans.pl (Downloads the HGNC gene families file using LWP:simple) ./convert_hgnc_gene_family_xlsx.pl (Not part of the generation of automated descriptions, but creates the excel spreadsheet for human gene families for viewing purposes) ./download_expression_cluster_summary_all_parallel.pl (downloads expression cluster data for all species)
Scripts for generating descriptions
- Creates sentences for tissue expression for C. elegans
- Creates sentences for non-elegans species for the orthologous elegans gene and the processes in which it's involved.
- this is a wrapper, runs several scripts, creates the orthology sentences for all species including elegans, reads the popularity file.
- Creates the sentences for process, function and sub-cellular localization for all species including elegans.
- Concatenates all sentences for the different semantic categories for all species
- Generates the file for import of the automated descriptions into the OA.
To create the concise descriptions run:
The new wrapper written by Yuling, for WS254 (April 2016), that calls all nine of the below scripts, run using sudo:
('space' between script names and the ampersands, do not introduce new line characters)
./create_sentence_tissue_expressions_elegans.pl && ./create_GO_sentences_elegans_species_parallel_all.pl && ./create_sentence_multiple_orthologs_species_all_parallel_all.pl && ./create_GO_sentences_species_parallel_all.pl && ./create_gene_regulation_expression_cluster_sentences_species_parallel_all.pl && ./create_molecule_regulation_expression_cluster_sentences_species_parallel_all.pl && ./create_anatomy_expression_cluster_sentences_species_parallel_all.pl && ./concatenate_sentences_species_parallel_all.pl && ./generate_OA_concise_descriptions_parallel_all.pl
Takes about 4.5 hours to run
Script for a separate file for manually curated genes
One can create a file named 'WBGenes_descriptions_for_manual.txt separately, to have the automated descriptions for only those genes that have been curated manually.
Script for Numbers
Finally, a report is written to detail the number of concise descriptions for each species for a given production release:
Moving files to Textpresso-dev for production
http://textpresso-dev.caltech.edu/concise_descriptions/ shows /home/ranjana/automated_descriptions_data which is actually a soft link to /data2/srv/textpresso-dev.caltech.edu/www/docroot/concise_descriptions
1. On firstname.lastname@example.org, go to :
concise_descriptions_data/release For the WS254 release folder, used r sync, so:
ranjana@brahma:~/concise_descriptions_data/release$ rsync -r WS254 email@example.com:/home/ranjana/data/.
gives this type of error message:
skipping non-regular file "WS254/b_malayi/orthology/input_files/b_malayi.orthologs.txt" for all 9 species,
Takes a couple of minutes, less than 5.
Previously used scp to do the same: then use 'scp -r' to scp the entire release directory, eg. WS253 to firstname.lastname@example.org:/home/ranjana/data
This takes about an hour to 1.5 hrs.
2. logon to Textpresso-dev,
then move (mv) WS253 from /home/ranjana/data to /home/ranjana/automated_descriptions_data/release
In: /home/ranjana/automated_descriptions_data/release mv ~/data/WS253/ .
Changes to software, files, workflow
June 2016: WS255
- Biomart main web services was still down, we used mirror site as recommended, changes in biomartWebExample.pl: "http://uswest.ensembl.org/biomart/martservice"
- Added a wrapper to get all source files: get_source_files_wrapper.sh, totally 26 scripts in it.
- For future runs, both "production_release.txt" and "release.txt" need to be changed, where production_release = release + 1
April 2016: WS254
- Biomart web services was down, so used the WS253 HumanIDs_mart_export.org.txt, protocol for Biomart web services may have changed, so check in advance for WS255.
- New wrapper written to run all the create* scripts, 9 scripts, called create_sentences_wrapper.sh
- Yuling modified ./create_sentence_tissue_expressions_elegans.pl, due to changes in anatomy_association_WS253
- will look for IDA instead of IEP
- will ignore all lines with the 'Enrich' qualifier, column 4, as this is the expression cluster data
Finding gene popularity (number of mentions in papers)
If one wants to create a listing of Textpresso publications associated with genes in either the 'results' section or the 'body' (if the paper can not be sectioned), then one would execute the following:
$ ./publication_paper_gene_results.pl $ ./publication_paper_gene_body.pl
If one wants to examine gene popularity for C. elegans in the literature, one can type:
$ Rscript textpresso_gene_popularity.R WS253
The output will be a PDF file named, textpresso_gene_popularity.pdf.
If one would like to create a flow chart of the perl scripts available in this respository:
$ ./graph_perl.pl ./my_perl_script.pl jpg ./my_output_directory
This script will produce a text file (my_perl_script.txt), a HTML file (my_perl_script.html) and (if there's enough memory available) a jpg file (my_perl_script.jpg) outlining the flow of data within the script. These files are written into the ./my_output_directory Other formats besides jpg are available: bmp canon dot gv xdot xdot1.2 xdot1.4 cgimage cmap eps exr fig gd gd2 gif gtk ico imap cmapx imap_np cmapx_np ismap jp2 jpg jpeg jpe pct pict pdf pic plain plain-ext png pov ps ps2 psd sgi svg svgz tga tif tiff tk vml vmlz vrml wbmp webp xlib x11
Software configuration management
Pushing files to GitHub:
copy any new developed files or scripts from /home/ranjana/concise_descriptions_scripts to /home/ranjana/latest_version_scripts:
$ cp /home/ranjana/concise_descriptions_scripts/README.md /home/ranjana/latest_version_scripts $ cd /home/ranjana/latest_version_scripts
git is installed and updated on brahma.textpresso.org and textpresso-dev.caltech.edu:
$ which git $ git --version $ git init (done once) $ git remote add origin https://github.com/WormBase/automated_descriptions.git (only done to change github repository) $ git add README.md $ git commit -m “Add Readme” $ git push
To check if your changes were uploaded, open a browser with URL, https://github.com/WormBase/automated_descriptions
To examine the number of lines of code (without comments or spaces), use cloc or sloccount:
$ cloc --version $ sloccount --version $ cloc /media/data1/jdone/concise_descriptions/concise_descriptions/automated_descriptions/automated_descriptions $ sloccount /media/data1/jdone/concise_descriptions/concise_descriptions/automated_descriptions/automated_descriptions