Predicting Transcriptional Responses to Unseen Compounds through Context Modeling and Multi-View Transfer
- 1Mohamed bin Zayed University of Artificial Intelligence
Abstract
Predicting drug-induced transcriptional responses could support in silico screening of new compounds, yet existing models barely use the compound: on a drug-blind L1000 benchmark, removing the drug input from seven published models changes accuracy by at most 0.012 Pearson correlation. They capture shared cellular and experimental variation while overlooking the compound-specific signals that distinguish drug effects. We present UnDetect, a framework that separates contextual variation from compound effects to predict responses to transcriptionally unprofiled compounds. UnDetect integrates chemical structure, annotated targets, mechanisms of action and Cell Painting morphology into a normalized multi-view kernel that accommodates incomplete compound information, and transfers residual responses within and across profiled cell lines, so that pharmacological and phenotypic relationships inform predictions beyond chemical similarity. On a drug-blind benchmark, where the paired control reveals the assay plate, UnDetect is the only model to beat a network that ignores the compound, in every fold; on a within-plate score that is zero for any compound-blind prediction, it reaches 0.117 against 0.092 for the strongest published model. On DeepCE's benchmark it reaches 0.602 Pearson correlation against the reported 0.491. Transfer across cell lines improves 28 of 28 LINCS L1000 and, at the two higher doses, 47 and 49 of 50 Tahoe-100M cell-line evaluations. These results show that separating context from compound effects and integrating complementary compound information improves transcriptional prediction and extends the utility of perturbation atlases for virtual screening.
Related
- Characterization and Behavioral Responses of Volatile Organic Compounds in Commercial Cockroach Baits
- Predicting and finding geroprotective compounds through modulating conserved longevity–associated aging modules
- Multi-omics dissection of transcriptional and post-transcriptional responses in cyanobacterial high-light adaptation
- Exogenous Melatonin Alleviates High- and Low-Temperature Stress in Cucumber Seedlings and Regulates Hormone Signaling-Related Transcriptional Responses
- Diaphorin induces extensive transcriptional remodeling of translation, transport, metabolism, and stress responses in Escherichia coli
BibTeX
@misc{lei2026predicting,
title = {{Predicting Transcriptional Responses to Unseen Compounds through Context Modeling and Multi-View Transfer}},
author = {Y. Lei and H. Liang and K. Peng and J. Wen},
howpublished = {bioRxiv},
year = {2026},
month = nov,
doi = {10.64898/2026.11.05.764505},
url = {https://www.biorxiv.org/content/10.64898/2026.11.05.764505v1},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - Predicting Transcriptional Responses to Unseen Compounds through Context Modeling and Multi-View Transfer
AU - Y. Lei
AU - H. Liang
AU - K. Peng
AU - J. Wen
PY - 2026
DA - 2026/11/05
PB - bioRxiv
DO - 10.64898/2026.11.05.764505
UR - https://www.biorxiv.org/content/10.64898/2026.11.05.764505v1
AB - Predicting drug-induced transcriptional responses could support in silico screening of new compounds, yet existing models barely use the compound: on a drug-blind L1000 benchmark, removing the drug input from seven published models changes accuracy by at most 0.012 Pearson correlation. They capture shared cellular and experimental variation while overlooking the compound-specific signals that distinguish drug effects. We present UnDetect, a framework that separates contextual variation from compound effects to predict responses to transcriptionally unprofiled compounds. UnDetect integrates chemical structure, annotated targets, mechanisms of action and Cell Painting morphology into a normalized multi-view kernel that accommodates incomplete compound information, and transfers residual responses within and across profiled cell lines, so that pharmacological and phenotypic relationships inform predictions beyond chemical similarity. On a drug-blind benchmark, where the paired control reveals the assay plate, UnDetect is the only model to beat a network that ignores the compound, in every fold; on a within-plate score that is zero for any compound-blind prediction, it reaches 0.117 against 0.092 for the strongest published model. On DeepCE's benchmark it reaches 0.602 Pearson correlation against the reported 0.491. Transfer across cell lines improves 28 of 28 LINCS L1000 and, at the two higher doses, 47 and 49 of 50 Tahoe-100M cell-line evaluations. These results show that separating context from compound effects and integrating complementary compound information improves transcriptional prediction and extends the utility of perturbation atlases for virtual screening.
ER -