Location Tests with Noisy Proxies for Latent Variables

TL;DR

This paper analyzes how noisy proxies of latent variables affect statistical tests and proposes adaptive methods to improve inference power, especially in single-cell CRISPR screens, by quantifying proxy quality and adjusting testing strategies.

Contribution

It introduces a framework to measure proxy quality via Pitman efficiency and develops adaptive tests that outperform existing methods regardless of proxy accuracy.

Findings

Weighted tests outperform unweighted when proxies are sufficiently accurate.

Proposed adaptive methods improve power across various proxy qualities.

Numerical simulations validate the effectiveness of the adaptive strategies.

Abstract

We investigate inference in a latent binary variable model where a noisy proxy of the latent variable is available, motivated by the variable perturbation effectiveness problem in single-cell CRISPR screens. The baseline approach is to ignore the perturbation effectiveness problem, while a recent proposal employs a weighted average based on the proxies. Our main goals are to determine how accurate the proxies must be in order for a weighted test to gain power over the unweighted baseline, and to develop tests that are powerful regardless of the accuracy of the proxies. To address the first goal, we compute the Pitman relative efficiency of the weighted test relative to the unweighted test, yielding an interpretable quantification of proxy quality that drives the power of the weighted test. To address the second goal, we propose two strategies. First, we propose a maximum-likelihood based approach that adapts the proxies to the data. Second, we propose an estimator of the Pitman efficiency if a "positive control outcome variable" is available (as is often the case in single-cell CRISPR screens), which facilitates an adaptive choice of whether to use the proxies at all. Our numerical simulations support the Pitman efficiency as the key quantity for determining whether the weighted test gains power over the baseline, and demonstrate that the two proposed adaptive tests can improve on both existing approaches across a range of proxy qualities.