Predictive Models of Prenatal Developmental Toxicity from ToxCast High-Throughput Screening Data
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Environmental Protection Agency's ToxCast project is profiling the in vitro bioactivity of chemicals to assess pathway-level and cell-based signatures that correlate with observed in vivo toxicity. We hypothesized that developmental toxicity in guideline animal studies captured in the ToxRefDB database would correlate with cell-based and cell-free in vitro high-throughput screening (HTS) data to reveal meaningful mechanistic relationships and provide models identifying chemicals with the potential to cause developmental toxicity. To test this hypothesis, we built statistical associations based on HTS and in vivo developmental toxicity data from ToxRefDB. Univariate associations were used to filter HTS assays based on statistical correlation with distinct in vivo endpoint. This revealed 423 total associations with distinctly different patterns for rat (301 associations) and rabbit (122 associations) across multiple HTS assay platforms. From these associations, linear discriminant analysis with cross-validation was used to build the models. Species-specific models of predicted developmental toxicity revealed strong balanced accuracy (> 70%) and unique correlations between assay targets such as transforming growth factor beta, retinoic acid receptor, and G-protein-coupled receptor signaling in the rat and inflammatory signals, such as interleukins (IL) (IL1a and IL8) and chemokines (CCL2), in the rabbit. Species-specific toxicity endpoints were associated with one another through common Gene Ontology biological processes, such as cleft palate to urogenital defects through placenta and embryonic development. This work indicates the utility of HTS assays for developing pathway-level models predictive of developmental toxicity.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it