P2P: Cross-View Population Denoising for Unpaired Single-Cell Perturbation Response Prediction

2026-09-28

AI Summary

This study addresses the challenges of unpaired data and individual noise interference in single-cell perturbation prediction by proposing P2P, a framework designed to extract reproducible population-level effects. P2P innovatively adopts a set-level supervision perspective that treats random cell sets as fundamental units. It employs a permutation-invariant encoder integrated with a gated residual network to incorporate empirical memory, alongside a cross-view denoising mechanism that effectively disentangles population responses from view-specific noise. Furthermore, structured perturbation tokens and a heteroscedastic prediction head are introduced to enhance modeling precision. Evaluated across four benchmark datasets, P2P achieves the lowest RMSE and highest Effect Pearson correlation, significantly outperforming baseline models such as GenePert.

Abstract

AIVC (AI Virtual Cell) is a learned simulator of cellular behavior across conditions. Predicting how a cell population responds transcriptionally to a genetic perturbation is a core task. Perturb-seq records that response by destructive sequencing, so a control cell and a perturbed cell are never observed as a pair, and cells under one condition remain heterogeneous and noisy. Regression on individual cells absorbs sampling variation into the estimated effect, whereas interpretation requires the reproducible population effect. P2P (Perturbation-to-Perturbation) takes a stochastic cell-set view as its supervision unit. Two views drawn from the same condition share a reproducible population effect and differ by view-specific variation. A permutation-invariant set encoder summarizes the control population, a structured encoder represents perturbation tokens, cellular context, dose, and combination interactions, and a gate blends empirical condition-effect memory with a neural residual. A heteroscedastic head predicts the population mean and gene-wise response variance. Under one protocol and five seeds, P2P attains the lowest expression RMSE and the highest Effect Pearson, DEG F1, and DEG average precision on each of Adamson, Norman, Replogle K562, and Replogle RPE1 relative to GenePert, LinearPert, SLIM, Scouter, and scPILOT. On Replogle K562, Effect Pearson rises from 0.643 to 0.702 and DEG F1 rises from 0.067 to 0.178 relative to Scouter, the strongest baseline on both metrics.

Problem

• single-cell perturbation response prediction
• unpaired data
• population denoising
• transcriptional response
• cell heterogeneity

Innovation

• Cross-View Population Denoising
• Permutation-Invariant Set Encoder
• Heteroscedastic Head
• Single-Cell Perturbation Prediction
• AI Virtual Cell