SicknessMiner

SicknessMiner extracts disease-disease associations (DDAs) from unstructured biomedical text using deep-learning-driven text-mining to centralize DDAs relevant to blood cancers (BCs) and other diseases.

Key Features:

• Centralization of Disease-Disease Associations: Aggregates DDAs from unstructured biomedical text and disparate sources into a centralized dataset.
• Deep-learning-driven text-mining: Applies deep-learning algorithms to process and analyze large volumes of raw biomedical text data.
• Named Entity Recognition (NER) and Normalization (NEN): Performs NER to identify disease entities and NEN to standardize them to common identifiers.
• Co-mention and similarity-based retrieval: Extracts DDAs via co-mention analysis and gene- or variant-disease similarity comparisons using DisGeNET data.
• High retrieval accuracy: Recovers approximately 92% of associations present in DisGeNET while identifying nearly 15% of DDAs unique to its pipeline.
• Focus on blood cancers (BCs): Targets extraction and analysis of associations particularly relevant to blood cancers (BCs) alongside other diseases.

Scientific Applications:

• Understanding Disease Pathogenesis: Enables exploration of links between diseases to inform hypotheses about shared mechanisms and pathways.
• Drug Repurposing and Development: Identifies shared molecular targets across diseases to inform drug repurposing strategies and target discovery.
• Personalized Medicine: Supports development of personalized treatment strategies by revealing a patient's disease association profile.

Methodology:

Preprocess raw biomedical text data; perform Named Entity Recognition (NER); perform Named Entity Normalization (NEN); extract disease-disease associations using co-mention strategies and gene- or variant-disease similarity analyses with DisGeNET; validate and compare extracted associations against DisGeNET.