Skip to main content

e(BE:L) - validation and extension of BEL networks.

Project description

e(BE:L) is a Python package built for both validating and modeling information extracted from publications using Biological Expression Language (BEL). This software package serves a comprehensive tool for all of your BEL needs and serves to create enriched knowledge graphs for developing and testing new theories and hypotheses.

e(BE:L) have implemented several other knowledge bases to extend the BEL knowledge graph or map identifiers.

Installation Current version on PyPI MIT

The easiest way to install ebel is to use docker-compose. See below instructions to use the docker installation.

ebel can be directly installed from PyPi with pip:

$ pip install ebel

But we want to encourage you to use the latest development version which can be installed with:

$ pip install git+https://github.com/e-bel/ebel

Package Requirements

Installing OrientDB

This software package is designed to work in conjunction with OrientDB, a NoSQL, multi-model database that acts as both a graph and relational database. e(BE:L) uses OrientDB for generating the knowledge graph derived from BEL files. To get started with e(BE:L), first download OrientDB and get a server up and running. The first time the server is ran, you will need to create a root password. Once it is up and running, you can get start importing BEL files into it!

On Linux you can use following commands:

wget https://repo1.maven.org/maven2/com/orientechnologies/orientdb-community/3.2.2/orientdb-community-3.2.2.tar.gz
tar -xvzf orientdb-community-3.2.2.tar.gz
cd orientdb-community-3.2.2/bin
./server.sh

SQL Databases

This package is capable of enriching the compiled knowledge graphs with a lot of external information, however, this requires a SQL databases for storage. While, a SQLite database can be used, this is not recommended as the amount of data and complexity of queries will be quite slow. Additionally, SQLite will not be directly supported, the methods will be built such that they should work with both SQLite and MySQL, but we will not address performance issues due to using SQLite.

Instead, we recommend setting up a MySQL server or MariaDB to use with e(BE:L). By default, PyMySQL is installed as a driver by e(BE:L), but others can also be used.

On Lunux Ubuntu you can use following command:

sudo apt install mysql-server -y

or:

sudo apt install mariadb-server -y

Configuration

Before you start working with e(BE:L), a simple to use wizard helps you to setup all configurations. Make sure OrientDB and MySQL (or MariaDB) are running. Then start the configuration wizard with:

ebel settings

The wizard will create the needed databases and users in OrientDB and MySQL/MariaDB.

Package Components

To test the different components you find here several BEL and already compiled JSON files.

BEL Validation

BEL is a domain-specific language designed to capture biological relationships in a computer- and human-readable format. The rules governing BEL statement generation can be quite complex and often mistakes are made during curation. e(BE:L) includes a grammar and syntax checker that reads through given BEL files and validates whether each statement satisfies the guidelines provided by BEL.bio. Should any BEL statement within the file not adhere to the rules, a report file is created by e(BE:L) explaining the error and offering suggested fixes.

You can use the following command to validate your BEL file:

$ ebel validate /path/to/bel_file.bel

In a single command, you can validate your BEL file as well as generate error reports if there are errors and if there are none, produce an importable JSON file:

$ ebel validate /path/to/bel_file.bel -r error_report.xlsx -j

BEL documents should be properly formatted prior to validation. e(BE:L) contains a repair tool that will check the format and it is highly recommended that this is used prior to validation. The repaired will overwrite the original if a new file path is not specified. Here is an example:

$ ebel repair /path/to/bel_file.bel -n /path/to/repaired_file.bel

Import Process

BEL Modeling - OrientDB

BEL files that have passed the validation process can be imported into the database individually or en masse. During the import process, e(BE:L) automatically creates all of the relevant nodes and edges as described in the BEL files. Additionally, e(BE:L) also automatically adds in missing nodes and edges that are known to exist e.g. protein nodes with a respective RNA or gene node with have these automatically added to the graph with the appropriate translatedTo and transcribedTo edges.

Model Enrichment - MySQL

e(BE:L) goes one step farther when compiling your BEL statements into a knowledge graph by supplementing your new graph model with information derived from several publicly available repositories. Data is automatically downloaded from several useful sites including UniProt , Ensembl, and IntAct and added as generic tables in your newly built database. Information from these popular repositories are then linked to the nodes and edges residing in your graph model, allowing for more complex and useful queries to be made against your data. This data is automatically downloaded, parsed, and imported into a specified SQL database.

Importing - Getting Started

e(BE:L) supports OrientDB as graph database and MySQL and MariaDB as RDBMS

Make sure you have downloaded/installed and running

  1. OrientDB

  2. MySQL or MariaDB
    1. MySQL
    2. MariaDB

This can be configured as a service in both Windows and Unix systems.

Set your MySQL connection parameters in e(BE:L):

$ ebel set-mysql --host localhost --user root --password myPassWord --database ebel

Once you have made sure both OrientDB and MySQL are running, you can now import an e(BE:L) compiled JSON file:

$ ebel import-json /path/to/checked_bel.json -u root -p orientdbPassword -d ebel -h localhost -p 2424

After you have successfully connected to the OrientDB database at least once, the login credentials will be written to the config file and no longer need to be passed (same with enrich command):

$ ebel import-json /path/to/checked_bel.json

You can also import all e(BE:L) compiled JSON files in a passed directory:

$ ebel import-json /path/to/bel_json/dir/

If you do no wish to enrich the graph, or wish to disable the protein/RNA/gene extension step, you can toggle these with the following options:

$ ebel import-json /path/to/checked_bel.json -e -g

You can run an enrichment step later using the enrich command:

$ ebel enrich

This command can also be given a list of resources to either skip or include during enrichment:

$ ebel enrich -i uniprot,hgnc

or:

$ ebel enrich -s intact,kegg

Querying the Graph

Once the knowledge graph has been created (and optionally enriched), one can use e(BE:L) to query it directly.

from ebel import Bel
bel = Bel()

relation_query = "SELECT out.bel as subject, @class as relation, in.bel as object FROM bel_relation"
df = bel.query(relation_query)  # Returns results as pandas DataFrame
result_list = bel.query_get_dict(relation_query)  # Returns results as a list of dicts

To specify a different graph database to connect to than the one in your config file, you can pass connection parameters when initializing the Bel class:

from ebel import Bel
config_params = {
    "db": "my_db",
    "user": "superdude",
    "password": "password123",
    "server": "localhost",  # defaults to localhost if none given
    "port": "2424",  # defaults to 2424 if none given
}

bel = Bel(graph_config=config_params)

# To overwrite your default values in the config file
bel = Bel(graph_config=config_params, overwrite_config=True)

Docker installation

Make sure docker and docker-compose are installed.

docker-compose up --build -d
docker exec -it ebel_ebel ebel settings

Several question will follow. You can accept the default values (just press RETURN) except the following questions:

OrientDB server [localhost] ?
ebel_orientdb
OrientDB root password (to create database and users)
ebel
MySQL/MariaDB sever name [localhost]
ebel_mysql
MySQL root password (will be not stored) to create database and user
ebel

It’s strongly recommended, if you are using ebel in the production environment, to change the standard root MySQL and OrientDB passwords in the docker-compose.yml file.

To load example files in container and import.

docker exec -it ebel_ebel git clone https://github.com/e-bel/example_json_bel_files.git
docker exec -it ebel_ebel ebel import-json example_json_bel_files/phago.bel.json -e

To enrich the network:

docker exec -it ebel_ebel ebel enrich

Following services are now available:

  1. OrientDB Studio

  2. e(BE:L) REST server

  3. phpMyAdmin

API

Finally, this package comes equipped with a built-in RESTful API using Flask. Users that have a running and populated set of databases can also create a running API server which contains several queries for retrieving information from both the network itself, as well as the downloaded enrichment information stored in the SQL database.

This server can be activated using:

$ ebel serve

You can also specify certain parameters as options:

$ ebel serve -p 5000 --debug-mode

Disclaimer

e(BE:L) is a scientific software that has been developed in an academic capacity, and thus comes with no warranty or guarantee of maintenance, support, or back-up of data.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ebel-1.0.37.tar.gz (503.9 kB view hashes)

Uploaded Source

Built Distribution

ebel-1.0.37-py3-none-any.whl (287.2 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page