back to Caltech documentation
Antibody curation SOPs
Antibody first pass
These are the important files for antibody curation
1. Antibody papers are identified via a script written by Juancarlos. Here is the first pass results for antibody curation: http://textpresso-dev.caltech.edu/azurebrd/wen/anti_protein_wen The file lists paper names and the antibodies associated with them. I download the file, save it and call it: anti_protein.txt
- save the file on desktop as 'anti_protein_20110113.txt' (add the date in file name)
- open a terminal
- cd to curation, then my_acefiles, then antibody_curation
- cp /Users/xiaodongwang/Desktop/anti_protein_20110113.txt . (copy the file into antibody_curation directory)
- cp anti_protein_20110113.txt anti_protein.txt (copy the contents in to anti_protein.txt as input file 1 when run TextpressoABFinder)
2. Curator maintain a curation log file for all the antibody papers that were already curated. The file is called: AbCurationLog.txt
3. The curation log file listed above only document papers that were curated after Texpresso first pass was applied. Antibodies curated before that are kept in this file: WBAbPaperList.ace
4. There is a script written by Wen to screen file 1, and filter out papers in file 2 and 3 (which were already curated), then give the new paper list. The script is called: TextpressoAbFinder.pl
Here is how to use the TextpressoAbFinder.pl
This script check the result of Textpresso, compare with the antibody paper list dumped from citace, and look For Antibody papers that were not curated.
Input file 1: anti_protein.txt -- all antibody papers found by Textpresso
Input file 2: WBAbPaperList.ace -- Antibody papers curated before Textpresso time
Input file 3: CurationLog/AbCurationLog.txt -- Antibody curation log.
Output file 1: NewAbPaper.txt -- New antibody papers (I can change the output file 1 name to NewAbPaper_20110113 in script)
Output file 2: TPAbFalsePositive.txt -- All false positive antibody papers.
1789 papers flagged by Textpresso, 1734 curated, 55 need to be checked. Among not curated papers, 30 has anti-XXX pattern, 25 has no anti-XXX pattern. 1626 papers curated in citace, 1347 found by Textpresso, 279 not found by Textpresso. Recall is 0.828413284132841. 518 papers identified by Textpresso are false positive. Precision is 0.710452766908888.
5. The result of the script is NewAbPaper.txt. This is the list of antibody papers that need to be curated.
Curation is done via Phenote or OA.
Curators need to document the status of every paper from NewAbPaper.txt into the curation log file AbCurationLog.txt, so that the same paper will not appear again next time.
Antibody curation controlled vocabulary
Antibody control vocabulary
Remark "Commercial Antibody." Remark "Tissue Specific Antibody Marker." Summary "Rabbit polyclonal antibody against XXX recombinant protein." Summary "Rabbit polyclonal peptide antibody against XXX." Summary "Mouse monoclonal peptide antibody against XXX."
Antibody curation guideline
WormBase requires the following information for Antibody:
1. Antibody Name: for consistance, use [WBPaperID]:anti-genename (_1, _2, etc, if several antibodies are made for same gene. genename is in CAPITALS. ig. [WBPaper00036348]:anti-EPG-2)
2. Original reference where the antibody was first reported. For antibodies that are published for the first time, list the original publication and mark the antibody as "Original_publication" antibody (these are good and valid antibody objects in WormBase.)
3. targeting gene (abc-1, xyz-1 ...), clonality (polyclonal or monoclonal) and animal (rabbit or mouse ...)
4. Antigen used to generate antibody (peptide or protein sequence)
5. If the antibody is from another paper, find the original antibody object and add the reference to it.
6. If the antibody has no original reference, create a new antibody object and mark it as "No_original_reference". If you suspect the antibody is the same as another one that was previously published, enter the "Possible_pseudonym" field.
is located in tazendra: /home/acedb/wen/phenote-antibody/dump_antibody_ace.pl
dump out file: ./dump_antibody_ace.pl
file name: antibody.ace
I usually change the file name: cp antibody.ace antibody.ace.date_of_dump
then copy file to spica: scp antibody.ace.20110503 email@example.com:/home/citace/Data_for_citace/Data_from_Xiaodong/.
Changed the postgres tables for ---05/22/2011
reference -> paper
location -> laboratory
email from Juancarlose related to cronjob --- 06/06/2011
Set the cronjob to run every Thursday : 0 2 * * thu /home/acedb/xiaodong/oa_antibody_dumper/dump_antibody_ace.pl It puts the file at :
So you can see it at :
If you need to run it manually, just paste into the shell :
Then log onto spica, cd into the directory where you want it, remove the existing antibody.ace file, and do :