You are here:
Atomic contribution mapping and exploration with reverse fingerprinting (ACME-RF): Assigning toxicological endpoints to chemical structure at atomic resolution (SOT 2021)
Citation:
Goldsmith, M., D. Chang, AND C. Williams. Atomic contribution mapping and exploration with reverse fingerprinting (ACME-RF): Assigning toxicological endpoints to chemical structure at atomic resolution (SOT 2021). Society of Toxicology 2021 Annual meeting, Virtual, NC, March 12 - 26, 2021. https://doi.org/10.23645/epacomptox.14474031
Impact/Purpose:
Poster presented to the Society of Toxicology annual meeting March 2021
Description:
The tools and model systems (in silico, invitro, in vivo) for assessing biological activity, both adverse (toxicological) or beneficial (pharmacological) are usually focused on diagnostic endpoints and metrics that are determinate, retrospective, reactionary and quantitative (i.e. LC50, EC50, AC50, IC50 etc…). Although it is useful from a structure-activity relationship (SAR) perspective to use descriptor-based methods or fingerprint approaches to assign, classify or quantify newly tested (or newly created) chemicals based on these analyses, these approaches often characterize a chemical entity with a single quantitative endpoint, and provide limited or no information about chemical structural elements important to the endpoint. Furthermore, suggestions that identify prospective avenues or opportunities for improvement are typically absent. This leaves one with limited direction to “point the finger” on a molecule, as in a design paradigm. One approach, reverse-fingerprinting (RF), provides a useful marriage between any given discretized endpoint (molecular, cellular, phenotypic, etc…), and any feature-based molecular fingerprint (MACCS, GPIDAPH, SHAPES etc…). The method produces a quantitative and visual representation of atomic contribution to a biological endpoint mapped on to any given molecular structure (Williams C, Schreyer SK. Comb Chem High Throughput Screen. 2009 May;12(4):424-39. doi: 10.2174/138620709788167953. PMID: 19442069. ) In this work we introduce the concept of atomic contribution mapping and exploration (ACME) using the RF framework in a newly designed interface for use in the Molecular Operating Environment (MOE). Using publicly available datasets we systematically explore (I) beneficial-safe acaricides (II) toxicophores for the estrogen receptor and (III) chemical photostability/photolability of drugs. In the first example we demonstrate how one can rapidly identify a class of acaricide that is not-toxic to beneficials (apis melifera) while still being toxic to the varroa mite (varroa destructor) using very basic insecticide-class information as inputs. In another example we use the ToxCast NVS_NR_hER dataset (165/2645 actives/inactives : https://comptox.epa.gov/dashboard/assay_endpoints/NVS_NR_hER ) to build a RF model that was subsequently used to identify the toxicophore of estrogen receptor alpha that directly maps to known co-crystallized structures, implying utility of the RF approach in target deconvolution. Finally, in the third example we explore photostability half-lives based on publicly available data (Blum, Kristin M.. “Phototransformation of pharmaceuticals in the environment: Multivariate modeling and experimental determination of photolysis half-lives.” (2013).) and identify moieties of chemistry that lead to photodegradation. Using this RF mapping approach, we could identify and visualize the moieties of the molecule that give rise to (I) specific effect or apical endpoints across species (II) specific target interactions for biomolecular assays or (III) physicochemical molecular liabilities such as photostability. The RF method can be used not only to identify toxic chemicals, but also to identify critical toxicophore fragments (demonstrated by comparing ligand fragment scores with known toxicophores and known ligand/protein contacts). The ability to highlight both beneficial and detrimental groups in a congeneric series using these three cases and the RF methodology is examined by comparison to known SAR, and shows fundamental utility in both a hazard assessment as well as a discovery context to design out molecular liabilities. [ This abstract solely represents the views of the authors and not the view of the Agency.]
URLs/Downloads:
DOI: Atomic contribution mapping and exploration with reverse fingerprinting (ACME-RF): Assigning toxicological endpoints to chemical structure at atomic resolution (SOT 2021)
SOT_ACME_RF_POSTER_GOLDSMITH_02042021.PDF (PDF, NA pp, 1496.518 KB, about PDF)