The number of known compounds far outnumbers the number of compounds with reference NMR or MS spectra. The same is true with regard to the number of known compounds and the number of compounds that can be bought, acquired or isolated.  Even if we did have authentic reference standards for every known chemical, we would still be challenged with identifying the “unknown unknowns” – those compounds that are not yet known to science but which are commonly found in our samples.  Therefore the problem of identifying unknown chemicals is actually three problems: 1) an incomplete list of known compounds; 2) a lack of authentic reference compound data and 3) a lack of potential or biologically feasible structures. In this presentation I will describe our efforts to address these three problems.  First, I will describe our ongoing efforts to compile more complete lists of known or “expected” compounds that can be found in the environment or in biological systems.  Second, I will describe methods our team has been developing to accurately predict NMR and MS/MS spectra of chemicals based purely on their structures. These include the programs called NMRPred (for NMR) and CFM-ID (for MS/MS). I will provide some examples showing how well these programs perform. Third, I will describe our efforts to predict or generate biologically feasible structures using the program called BioTransformer.  Our hypothesis is that many of the “unknown unknowns” are compounds that have been chemically or biologically transformed via enzymatic or spontaneous reactions in the body or in the environment. BioTransformer is capable of predicting a wide range of biologically feasible structures through computer-simulated biotransformation of known structures. I will conclude my presentation with some speculative thoughts on where the field  of in silico metabolomics is going and what other opportunities exist to improve computer-aided compound identification.

Tools and databases for <i>in silico</i> compound identification

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.

Tools and databases for in silico compound identification