Response Magnitude as a Dominant Signal for Held-Out CRISPRi Perturbation Effect Prediction

Predicting the magnitude of a CRISPRi perturbation's transcriptomic effect on held-out target genes is an important open problem in single-cell biology. Recent work has documented that simple baselines often match or exceed deep perturbation predictors on related protocols. We study this phenomenon on the Virtual Cell Challenge (VCC) benchmark under a strict held-out target-gene split, identify the specific low-dimensional signal that drives the gap, and characterize how it transfers across cell types. The target is the log Anderson-Darling distance from non-targeting controls, which is strongly predictable from four deterministic scalar functions of the 2,000-dimensional input. A deep MLP encoder with direct access to the full input collapses toward the marginal training mean, and standard remedies do not close the gap. A linear regression on the four magnitude scalars alone exceeds the strongest x-only classical model, while a Random Forest on the input plus the four scalars substantially outperforms our deep proof-of-concept encoder. Two pre-specified controls attribute the magnitude gain to per-row alignment rather than added dimensionality. Under zero-shot transfer to two external CRISPRi screens evaluated against a target-gene endpoint rebuilt from single-cell data, magnitude-only predictors transfer positively whereas expression-only predictors are negative or unresolved. Exposing magnitude to the deep encoder improves transfer over its expression-only counterpart, yet the encoder does not outperform a four-scalar linear regression on the same features. We also find that the Anderson-Darling column distributed with these screens measures transcriptome-wide response breadth rather than target-gene effect strength, so evaluating transfer against it scores a different outcome.

Learning Stable Predictors from Weak Supervision under Distribution Shift

Learning from weak or proxy supervision is common when ground-truth labels are unavailable, yet robustness under distribution shift remains poorly understood, especially when the supervision mechanism itself changes. We formalize this as supervision drift, defined as changes in P(y | x, c) across contexts, and study it in CRISPR-Cas13d experiments where guide efficacy is inferred indirectly from RNA-seq responses. Using data from two human cell lines and multiple time points, we build a controlled non-IID benchmark with explicit domain and temporal shifts while keeping the weak-label construction fixed. Models achieve strong in-domain performance (ridge R^2 = 0.356, Spearman rho = 0.442) and partial cross-cell-line transfer (rho ~ 0.40). However, temporal transfer fails across all models, with negative R^2 and near-zero correlation (e.g., XGBoost R^2 = -0.155, rho = 0.056). Additional analyses confirm this pattern. Feature-label relationships remain stable across cell lines but change sharply over time, indicating that failures arise from supervision drift rather than model limitations. These findings highlight feature stability as a simple diagnostic for detecting non-transferability before deployment.