You are here:
Transcriptomics-Based Points of Departure for Fish – Use of In Silico Sub-Sampling to Inform High Throughput Assay Design
Citation:
Hazemi, M., K. Flynn, K. Vitense, M. Le, K. Bush, AND Dan Villeneuve. Transcriptomics-Based Points of Departure for Fish – Use of In Silico Sub-Sampling to Inform High Throughput Assay Design. SETAC North America 42nd Annual meeting, Portland, OR, November 14 - 18, 2021. https://doi.org/10.23645/epacomptox.17099990
Impact/Purpose:
Transcriptomics-based points of departure (POD) measured in short-term assays have shown as alternatives to long term testing in mammals. We are in the process of developing and evaluating the use of transcriptomics-based PODs for aquatic organisms for predicting chronic toxicity outcomes in longer term ecotoxicity tests. An important component of the assay development involves identifying both the amount of replication and overall gene coverage needed to generate reliable POD estimates. The present study performed in silico sub-sampling on 10 pilot datasets to inform experimental design optimization and assay acceptance criteria. This work will help ensure that subsequent research on these approaches is conducted in a manner that minimizes assay costs while maximizing the ability to generate quality data.
Description:
Aquatic organisms are highly vulnerable to chemical exposure and are therefore commonly used in environmental hazard assessment. Traditional animal toxicity testing requires extensive resources and time. Alternative test methods involving high-throughput transcriptomics are being evaluated to characterize the potency and potential hazards of chemicals to aquatic organisms. However, while costs of transcriptome sequencing have decreased significantly, there is still a need to balance robustness of the assay design against the cost of data generation. The current study utilized in silico subsampling to help inform selection of minimum sample and gene set sizes to incorporate into subsequent transcriptomics-based assays. Individual larval fathead minnows (Pimephales promelas; 1 day post-hatch) were exposed to a diverse set of chemicals (three metals, three pharmaceuticals, four pesticides) for 24 hours in 1-ml, 96-deep well plates with 12 concentrations per chemical and eight biological replicates per concentration. RNA was extracted from each larval fish sample following homogenization. Sequence data and the benchmark dose (BMD) modeling software BMDExpress2 were used to calculate transcriptomic-based points-of-departure (tPODs) that can be anchored to apical responses (e.g., mortality, altered behavior). Random, in silico subsampling of biological replicates and the transcriptome were performed to ascertain assay parameters that generated consistent point-of-departure estimates while minimizing the use of animals and overall sequencing costs. Results suggest that assay design should include at least 10,000 transcript features and five biological replicates. Our results also identify a need to develop acceptance criteria for potency estimates based upon assessment of the overall BMD distribution and a minimum number of differentially expressed genes. These results will inform the optimal design of subsequent high throughput transcriptomics assays with fathead minnow and other ecologically relevant aquatic organisms. Reliable generation of tPODs that have reasonable, and quantifiable, levels of uncertainty based on minimum assay performance criteria is expected to aid the adoption of these approaches for regulatory hazard characterization. This abstract neither constitutes nor necessarily reflects USEPA policy.
URLs/Downloads:
DOI: Transcriptomics-Based Points of Departure for Fish – Use of In Silico Sub-Sampling to Inform High Throughput Assay Design
HAZEMI_FHM TPOD - IN SILICO SUBSAMPLING TO INFORM ASSAY DESIGN_FINALSLIDES.PDF (PDF, NA pp, 2463.576 KB, about PDF)