gopreprocess

This repo holds parsing and GAF generation code for generating automatic annotations based on orthology, or based on upstream products from Protein2GO.

The Makefile helps coordinate the process for a single species, but each step can be run independently.

For example, running this preprocessing pipeline for MGI to create a single GAF 2.2 file that represents all the automated annotations that the GO Consortium will consume for mouse (this includes the human and rat annotations that can be converted to mouse via orthology, and the mouse annotations created by Protein2GO):

make run

make all will download the necessary upstream files like the Human & Human Isoform annotations, the Rat annotations, the MGI GPI file, the Alliance orthology file, the Mouse Xref file and the Gene Ontology in JSON format. It will then run the orthology annotation conversion step for each species, finally combining the resulting GAFs with the Protein2GO GAFs. The Protein2GO GAF files are left unchanged in this conversion process. The resulting GAFs (the MGI annotations created via orthology to Human, the MGI annotations created via orthology to Rat, and the Protein2GO annotations) are then concantenated together under a single header for use in the regular GO pipeline as an upstream MGI, automated annotation GAF.

To run just the download step:

make download_rat
make download_human

To run the conversion from rat or human annotations to mouse annotations:

make convert_rat
make convert_human

To run the download of the Protein2GO annotations:

make convert_p2g_annotations

to merge together the resulting GAFs:

make merge_gafs

to validate the resulting GAF:

make validate_merged_gafs 

This process is conducted by the silver-issue-325-gopreprocess GO pipeline repository branch's Jenkinsfile and the process is best summarized by the following stages:

1. The Download GAFs stage - this is where the orthologs species files and the Alliance orthology file are downloaded and is executed via the pipeline jenkins file which calls make download_rat and make download_human

2. The Generate automated annotations stage - this is where the orthologs species files are converted to mouse annotations and is executed via the pipeline jenkins file which calls make convert_rat and make convert_human. It also calls make convert_p2g_annotations to download the Protein2GO annotations. Finally, it calls make merge_gafs to merge the resulting GAFs together and make validate_merged_gafs to validate the resulting GAF. If more than 3% of the resulting merged GAF is invalid, the validate_merged_gafs target will return a non-zero exit code and the pipeline will fail.

run diff code to compare with an existing file

poetry run compare -file1 file_path -file2 file_path -o file_output_prefix

By default, this will compare subject_id (gene or gene product id), object_id (go term id), and evidence_code between the two files reporting the matching associations, the associations unique to file1 and those unique to file2.