This tool allows you to explore relationships between articles that have common accession numbers (e.g. uniprot, refseq, refsnp), and then export graph in JSON, GML and GXMML or save as SVG.
You can try it online on Google Cloud Platfrom http://35.246.2.161:9090/.
Sample Paper IDs to put in query: PMC6299211, 30510815, PMC6218573, 30425980, 30425978
Webpack is currently disabled.
For all of these to work you need Neo4j server.
https://neo4j.com/download-center/#releases
Python dependencies:
pandas, requests
How Python script works
- Download dump of accession numbers with associated paper indices from EuropePMC.
ftp://ftp.ebi.ac.uk/pub/databases/pmc/TextMinedTerms/ - Concatenate all csv files into one big table.
- Remove duplicates and create indices for unique values
- Add indices to all values (including duplicates) in big table
- Create relationship table
- Add titles to the table with papers
- Import everything into Neo4j using neo4j-admin import command
How to run everything:
- Specify paths in PATHS.txt and run main.py script. It takes up to 30 minutes to finish everything.
- Write in following settings in Neo4j config file (NEO4J_home_folder/conf/neo4j.conf):
dbms.active_database=imp.db (put name of your database here)
dbms.security.auth_enabled=false (if you don't want to set up a password)
- Open code of index.js and specify your login, password, and path to server; or leave them as they are (neo4j, 1234 localhost:7474)
- Run index.html
Warning:
Because of current limitatins of admin-import tool If you want to run script again, you have to delete the database "imp.db" or specify another name for it.
Ready database for Neo4j is available here.
Just put it in NEO4J_home_folder/data/databases.
Target specific search
Input
PMCID or PMID
Return
A graph of accession numbers that article of interest has and other articles that are connected to these accession numbers
Input
One accession number (with prefix 'acc:') acc:4H6B
A list of accession numbers: acc[4H6B,5BU3]
Return
A graph of articles that have these accession numbers
Input
One accession number and target databases: acc:4H6B AND db[pdb,ena]
Multiple accession numbers and target databases: acc[4H6B,5BU3] AND db[pdb,ena]
Return
A graph of accession numbers and articles that have accession numbers from specified databases
Check if two items of interest are somehow connected
Input
Paper and paper (PMCID or PMID): con(30425980, PMC6218573)
Paper and accession number: con(30425980, acc:3MWD)
Accession number and accession number: con(acc:3C4E acc:3MWD)
Return
A graph which shows a connection between two items up to 15 jumps
Free text search
Input
Text specifying boolean AND/OR connection, with '@' sign on the beginnig: @beta-catenin OR β-catenin
With limit for number of papers to retrieve: @beta-catenin OR β-catenin @ LIMIT 1000
Visualization script is modified version of tool made by @micwan88.