Biomedical articles share annotations with their citation neighborhood

Author: Raul Rodriguez-Esteban

Reference: Rodriguez-Esteban R. Biomedical articles share annotations with their citation neighbors. BMC Bioinformatics. 2021 Feb 26;22(1):95.

Paper link: https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-021-04044-4

The entire analysis can be run using the script run_all.sh, which contains the sequence of scripts that needs to be followed. The only thing missing in the run_all.sh script is the download of the MEDLINE, NLM2007, L1000 and BC2GN datasets.

Below you can find a brief description of some of the files in this repository.

Download data files:

• download_citations.sh: download citations from the Open Citation Index. The URL has to be adjusted for newer releases of the data.
• download_gene2pubmed.sh: download the gene2pubmed database
• download_pmid_doi_mapping.sh: download the PMID-PMC-DOI mappings from EBI
• download_uniprot.sh: download the UniProtKB annotations
• download_mesh_tree.sh: download MeSH tree structure

Processing data files:

• read_gene_annotations_uniprot.py

• read_gene_annotations_gene2pubmed.py

• read_mutation_annotations_dbsnp.py

• read_gene_annotations_bc2.py

• read_mesh_annotations.py

• mean_average_precision.py: main script computing recall, precision and MAP

For the BC2GN analysis:

• bc2gn_mean_average_precision.py: MAP values for BC2GN annotations
• bc2gn_recall.py: recall values for BC2GN annotations

Additional files:

• create_reference_annotations.py: processes the reference datasets NLM2007 and L1000
• create_shuffled_network.py: creates a shuffled version of the original network
• map_doi_citations_to_pmid.py: map citation pairs in DOI format to PMID using the EBI mapping data
• neighborhood_annotation_statistics_all_medline.py: network neighbor stats for all MEDLINE records
• parse_medline_for_mesh.py: list all MeSH major topic annotations in MEDLINE
• parse_medline_for_pmids.py: list all MEDLINE PMIDs