You are here:
Cross-species extrapolation of toxicity information using the Sequence Alignment to Predict Across Species Susceptibility (SeqAPASS) tool
Citation:
LaLone, C. Cross-species extrapolation of toxicity information using the Sequence Alignment to Predict Across Species Susceptibility (SeqAPASS) tool. International Fresenius: Endocrine Disruptors, Bonn, GERMANY, November 15 - 16, 2016.
Impact/Purpose:
not applicable
Description:
In the United States, the Endocrine Disruptor Screening Program (EDSP) was established to identify chemicals that may lead to adverse effects via perturbation of the endocrine system (i.e., estrogen, androgen, and thyroid hormone systems). In the mid-1990s the EDSP adopted a two tiered approach for screening chemicals that applied standardized in vitro and in vivo toxicity tests. The Tier 1 screening assays were designed to identify substances that have the potential of interacting with the endocrine system and Tier 2 testing was developed to identify adverse effects caused by the chemical, with documentation of dose-response relationships. While this tiered approach was effective in identifying possible endocrine disrupting chemicals, the cost and time to screen a single chemical was significant. Therefore, in 2012 the EDSP proposed a transition to make greater use of computational approaches (in silico) and high-throughput screening (HTS; in vitro) assays to more rapidly and cost-efficiently screen chemicals for endocrine activity. This transition from resource intensive, primarily in vivo, screening methods to more pathway-based approaches aligns with the simultaneously occurring transformation in toxicity testing termed “Toxicity Testing in the 21st Century” which shifts the focus to the disturbance of the biological pathway predictive of the observable toxic effects. An example of such screening tools include the US Environmental Protection Agency’s (EPA) ToxCast Program, consisting of a suite of ~650 HTS assays for automated and rapid identification of chemical bioactivities relevant to endocrine pathways and the adverse outcome pathway (AOP) framework which facilitates the use of these mechanistic or pathway-based data to link chemical perturbation to an adverse outcome of regulatory concern. As these tools are utilized to screen and identify chemicals with endocrine activity as a means to protect both humans and wildlife, there is a recognized challenge in understanding how broadly these data can be extrapolated to other species. Typically, HTS assays use mammalian model systems and AOPs are developed with a handful of species (or even one species) where toxicity data is available. Therefore, important knowledge gaps for extrapolating chemical effects across broad taxonomic groups exist. The US EPA’s Sequence Alignment to Predict Across Species Susceptibility (SeqAPASS; https://seqapass.epa.gov/seqapass/) tool, is a publically available online tool for evaluating molecular target conservation as a means to predict chemical susceptibility across species. With knowledge of the protein target(s) with which a chemical interacts to produce its effect and a known sensitive species, the SeqAPASS tool can be queried to compare millions of protein sequences from thousands of species to identify those most similar to the query sequence. This evaluation assumes that the more similar a chemical’s molecular target is to a known sensitive species, the more likely the chemical can interact with that similar protein in another species. Therefore, data from SeqAPASS provides a line of evidence to predict potential chemical susceptibility across species. Depending on available information about a chemical interaction with a protein target and the degree of protein characterization, the SeqAPASS evaluation can assess three levels of sequence similarity. Level 1 in the SeqAPASS evaluation compares primary amino acid sequences, providing susceptibility predictions with rather limited taxonomic resolution, likely identifying differences between vertebrates and invertebrates. Level 2 adds taxonomic resolution to the predictions by evaluating similarity between known functional domains within the protein (e.g., ligand binging domain). Level 3 allows for comparisons of individual amino acid residues across species to provide another line of ev