Science Inventory

Chemical curation of analytical method publications using the API-driven PubMed Abstract Sifter

Citation:

Baker, N., E. Carr, G. Janesch, AND A. Williams. Chemical curation of analytical method publications using the API-driven PubMed Abstract Sifter. ACS, New Orleans, LA, March 17 - 21, 2024. https://doi.org/10.23645/epacomptox.25994503

Impact/Purpose:

N/A

Description:

Analytical methods, specifically those focused on the identification of chemicals by liquid and gas chromatography mass spectrometry (LC/MS and GC/MS), are key approaches supporting targeted and non-targeted environmental analysis. Different chemicals and chemical classes are amenable to different spectrometry methods and chromatography conditions. The USEPA Center for Computational Toxicology and Exposure has developed a working prototype web search and retrieval application called AMOS (the Analytical Methods & Open Spectral database) hosting over 4000 analytical methods and over 200k associated mass spectra. The assembly of these data has included identification of method collections from various agencies (e.g., EPA, DEA and USGS) as well as the harvesting of data from peer-reviewed publications. The harvesting of chemical information embedded in the text of the chemical literature has been an essential aspect of this project by extracting chemical entities from publications and method documents to build up an underlying database of chemical(s) – document pair relationships. In this presentation we demonstrate how semi-automated curation work is made possible using the publicly available Abstract Sifter tool. We will cover how to query PubMed for relevant publications, rank the returned documents, and extract the chemical entities from the citation or the pdf, and finally, to associate the chemical name or CAS with its EPA DSSTox identifier. The resulting extracted data are a critical component of the AMOS database. This abstract does not necessarily represent US EPA policy.