GWAS Miner

Overview

GWAS Miner was created as part of my PhD project at the University of Leicester, tackling the problem of extracting meaningful data from GWAS publication text.

Features

• Extraction of genotype to phenotype associations, including genetic marker, disease and significance score (p-value).
• Visualisation of entity recognition and sentence structure within GWAS publication text.
• View publication statistics such as number of ontology disease term occurrences.

Running GWAS Miner

GWAS Miner can be utilised both with a graphical user interface and through passing command line parameters. Launching the graphical user interface can be done by passing the -g parameter to GWASMiner.py, allowing quick and easy access to all of it's features.

Input Files

GWAS Miner is designed to utilise BioC-JSON files such as those generated by the Auto-CORPus project (https://github.com/omicsNLP/Auto-CORPus), including the produced Tables-BioC JSON files.

Graphical User Interface

python GWASMiner.py -g

Command line

The following subset of features are available without launching the graphical user interface.

Process files within a directory

python GWASMiner.py -d <path_to_directory>

Update ontology cache

python GWASMiner.py -u

Visualise entities identified within a document

python GWASMiner.py -d <path_to_file> -g "ents"

Visualise sentence dependencies within a document

python GWASMiner.py -d <path_to_file> -g "sents"

Dependencies

The following Python packages are required to run GWAS Miner, using at least python3.5 or later:

• python-dateutil
• rdflib
• owlready2
• lxml
• jsonschema
• rtgo
• networkx
• spacy
• SciSpaCy pre-trained data model: pip install https://s3-us-west-2.amazonaws.com/ai2-s2-scispacy/releases/v0.2.5/en_core_sci_md-0.2.5.tar.gz
• svglib
• reportlab
• PyQt5

Contact

For issue reporting and feedback/recommendations please email Thomas Rowlands at [email protected].