PHAROS: turning single-cell perturbation models into target-directed drug-combination screens
- 1Stanford University
Abstract
Combination therapies are central to cancer treatment, but exhaustive screening is impractical. We introduce PHAROS, a framework that turns a pretrained single-cell perturbation model into a target-directed search engine for drug combinations. PHAROS predicts how a cell population changes under a drug, one drug at a time, then chains these predictions together to simulate drug combinations. It scores each simulated outcome against the desired target state and uses a search algorithm to find the most promising combinations, all without retraining the underlying model. Across two independent combinatorial perturbation datasets, PHAROS recovered exact or mechanism-matched two-drug responses in cell lines, both seen and unseen during model training. Its rankings were specific to the requested conversion and were not explained by single-drug effects, additive effects, or shared mechanism of action. In exploratory analyses of patient-derived metastatic HR+/HER2− breast tumors and basal cell carcinoma (BCC), PHAROS prioritized FDA-approved regimens, distinguished combinations by their predicted tumor-versus-immune objective profiles, and nominated pathway-level hypotheses, while explicitly identifying both tumor cohorts as outside the model's supported distribution. PHAROS provides a modular route from pretrained virtual-cell models to inverse, single-cell combination-screening platforms.
Related
- How many cells resolve a perturbation direction? A closed-form, control-aware cell quota for single-cell perturbation screens
- Plate design confounds donor variation and model evaluation in single-cell perturbation screens
- ISP Platform powered by Geneformer: Framework for Cross-Species, Sequential, and Multi-Gene In Silico Perturbation Screens with Application to iPS Cell State Transitions
- Lateralized dopaminergic neuron activity encodes goal-directed turning
- Pooled CRISPR screens identify genes and non-coding genomic regions that regulate red blood cell density
BibTeX
@misc{bezney2026pharos,
title = {{PHAROS: turning single-cell perturbation models into target-directed drug-combination screens}},
author = {J. Bezney and C. Ruggeri and F. Borra and L. S. Qi and F. M. Buffa and L. M. Steinmetz},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.08.749477},
url = {https://www.biorxiv.org/content/10.64898/2026.09.08.749477v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - PHAROS: turning single-cell perturbation models into target-directed drug-combination screens
AU - J. Bezney
AU - C. Ruggeri
AU - F. Borra
AU - L. S. Qi
AU - F. M. Buffa
AU - L. M. Steinmetz
PY - 2026
DA - 2026/09/08
PB - bioRxiv
DO - 10.64898/2026.09.08.749477
UR - https://www.biorxiv.org/content/10.64898/2026.09.08.749477v1
AB - Combination therapies are central to cancer treatment, but exhaustive screening is impractical. We introduce PHAROS, a framework that turns a pretrained single-cell perturbation model into a target-directed search engine for drug combinations. PHAROS predicts how a cell population changes under a drug, one drug at a time, then chains these predictions together to simulate drug combinations. It scores each simulated outcome against the desired target state and uses a search algorithm to find the most promising combinations, all without retraining the underlying model. Across two independent combinatorial perturbation datasets, PHAROS recovered exact or mechanism-matched two-drug responses in cell lines, both seen and unseen during model training. Its rankings were specific to the requested conversion and were not explained by single-drug effects, additive effects, or shared mechanism of action. In exploratory analyses of patient-derived metastatic HR+/HER2− breast tumors and basal cell carcinoma (BCC), PHAROS prioritized FDA-approved regimens, distinguished combinations by their predicted tumor-versus-immune objective profiles, and nominated pathway-level hypotheses, while explicitly identifying both tumor cohorts as outside the model's supported distribution. PHAROS provides a modular route from pretrained virtual-cell models to inverse, single-cell combination-screening platforms.
ER -