GeoBoost2

GeoBoost2 extracts and enriches GenBank metadata with geographic information about infected hosts to support virus phylogeography and genomic epidemiology.

Key Features:

• Metadata enrichment: Processes GenBank metadata accessions to identify and extract geographic locations of infected hosts.
• Natural language processing (NLP): Leverages NLP techniques to automate extraction of location data from nucleotide sequence repositories such as NCBI GenBank.
• Deep learning: Incorporates state-of-the-art deep learning techniques to improve the accuracy and reliability of location extraction from text data.
• Multi-source mining: Extracts geographic locations from PubMed and PMC-OA abstracts and from general text files.
• Performance optimization: Implements end-to-end extraction improvements that increase extraction performance and processing speed.
• Output for downstream analyses: Produces enriched geographic metadata suitable for phylogeographic and genomic epidemiology analyses.

Scientific Applications:

• Virus phylogeography: Provides host-location metadata to support reconstruction of viral geographic spread and phylogeographic inference.
• Genomic epidemiology: Enables incorporation of geographic metadata into genomic epidemiology analyses for tracking pathogen transmission.
• Pathogen surveillance: Supports surveillance activities within and across national borders by enriching sequence metadata with geographic information.

Methodology:

Applies an NLP pipeline and state-of-the-art deep learning techniques to process GenBank metadata accessions and extract geographic locations of infected hosts, and mines PubMed/PMC-OA abstracts and general text files for location mentions.