Stable-Shift: Biologically Structured Prediction of Transcriptional Responses to Unseen Gene Perturbations
Abstract
Predicting transcriptional responses to genetic perturbations could reduce the experimental burden of functional genomics, but extrapolation to genes that were never perturbed during training remains difficult. We present Stable-Shift, a structured method for estimating unseen-gene responses. Stable-Shift aggregates single-cell measurements into perturbation-level expression shifts, fits a low-rank response basis using training perturbations only, and predicts an unseen gene's coordinates in that basis from biological context. The context combines STRING interactions, network structure, control-cell expression statistics, and Gene Ontology annotations; the evaluated implementation uses graph convolution to integrate these inputs. On the supplied K562 Perturb-seq benchmark, Stable-Shift obtained 0.592 cosine similarity, compared with 0.569 for GEARS, together with higher Spearman correlation and top-gene precision among the evaluated methods. Its mean cosine similarity over five unseen-gene splits was 0.589 +/- 0.008. The same ordering was observed in the supplied graph-aware, residualized, gene-space, and Norman-dataset comparisons. These results support further study of biologically structured latent-response prediction, while the lower gene-space accuracy and sensitivity to sparse graph neighborhoods limit the scope of the present conclusions.
Versions
Related
- A conserved transcriptional axis links cortical organization, developmental timing and autism-associated perturbations
- Transcriptional perturbation effects remain structured beyond the leading global trend
- Biologically grounded locality priors close the data gap for vision transformers in neural prediction
- 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
BibTeX
@misc{dip2026stableshift,
title = {{Stable-Shift: Biologically Structured Prediction of Transcriptional Responses to Unseen Gene Perturbations}},
author = {Sajib Acharjee Dip and Liqing Zhang},
howpublished = {arXiv},
year = {2026},
month = jun,
doi = {10.48550/arxiv.2606.24940},
eprint = {2606.24940},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2606.24940},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Stable-Shift: Biologically Structured Prediction of Transcriptional Responses to Unseen Gene Perturbations
AU - Sajib Acharjee Dip
AU - Liqing Zhang
PY - 2026
DA - 2026/06/22
PB - arXiv
DO - 10.48550/arxiv.2606.24940
UR - https://arxiv.org/abs/2606.24940
AB - Predicting transcriptional responses to genetic perturbations could reduce the experimental burden of functional genomics, but extrapolation to genes that were never perturbed during training remains difficult. We present Stable-Shift, a structured method for estimating unseen-gene responses. Stable-Shift aggregates single-cell measurements into perturbation-level expression shifts, fits a low-rank response basis using training perturbations only, and predicts an unseen gene's coordinates in that basis from biological context. The context combines STRING interactions, network structure, control-cell expression statistics, and Gene Ontology annotations; the evaluated implementation uses graph convolution to integrate these inputs. On the supplied K562 Perturb-seq benchmark, Stable-Shift obtained 0.592 cosine similarity, compared with 0.569 for GEARS, together with higher Spearman correlation and top-gene precision among the evaluated methods. Its mean cosine similarity over five unseen-gene splits was 0.589 +/- 0.008. The same ordering was observed in the supplied graph-aware, residualized, gene-space, and Norman-dataset comparisons. These results support further study of biologically structured latent-response prediction, while the lower gene-space accuracy and sensitivity to sparse graph neighborhoods limit the scope of the present conclusions.
ER -