This is the adapter for the BioCypher representation of the Dependency
Map dataset. It assumes presence of the raw (v0.5) input data in the
data/v0.5/ folder (in the respective subfolders cellModels,
compounds, and genes). I made one change to remove a multi-line
field from cellModels_all.txt in line 832 (simply removing the
newline). The adapter uses the ontological mapping specified in
config/schema_config.yaml to assign input data onto the ontological
hierarchy supplied by Biolink (flexibly extended using various BioCypher
methods).
The adapter (adapter.py) currently only generates neo4j-admin import-ready files via the batch writer class (to biocypher-out or
any specified location). Interaction with the graph via the driver will
be implemented next. The process can be run using script.py. If all
goes well, the output folder will contain a neo4j-admin-import-call.sh
file which, when run in the terminal in the database folder, will create
a new database under the name specified in the db_name argument of the
BioCypherAdapter instantiation in script.py. The database does not
need to be running at this point (but it can). However, it is important
that the target database (in the current example the one named "import")
is either completely empty or does not exist yet.
After import, and in case the database in point did not exist yet, the
database can be created and activated in the Neo4j browser with :use system, create database <db_name>, :use <db_name>. At this point,
the DBMS needs to be running.
The project can be installed using poetry:
git clone https://github.com/saezlab/BioCypher
git clone https://github.com/saezlab/DepMap-BioCypher
cd DepMap-BioCypher
poetry install
The installation assumes a clone of the BioCypher repo to be placed in
the same parent folder (e.g., /User/GitHub), which is the reason for
the first line.