Assumption-Lean Post-Integrated Inference with Surrogate Control Outcomes
Not on the map. Why?
Abstract (arXiv)
Data integration methods aim to extract low-dimensional embeddings from high-dimensional outcomes to remove unwanted variations, such as batch effects and unmeasured covariates, across heterogeneous datasets. However, multiple hypothesis testing after integration can be biased due to data-dependent processes. We introduce a robust post-integrated inference method that accounts for latent heterogeneity by utilizing control outcomes. Leveraging causal interpretations, we derive nonparametric identifiability of the direct effects using negative control outcomes. By utilizing surrogate control outcomes as an extension of negative control outcomes, we develop semiparametric inference on projected direct effect estimands, accounting for hidden mediators, confounders, and moderators. These estimands remain statistically meaningful under model misspecifications and with error-prone embeddings. We provide bias quantifications and finite-sample linear expansions with uniform concentration bounds. The proposed doubly robust estimators are consistent and efficient under minimal assumptions and potential misspecification, facilitating data-adaptive estimation with machine learning algorithms. Our proposal is evaluated using random forests through simulations and analysis of single-cell CRISPR perturbed datasets, which may contain potential unmeasured confounders.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 22 of its 67 references that are in the map. First posted on arXiv 7 Oct 2024.
Builds on
22 of 67 references.
References · 22
- 2016Identifying Causal Effects With Proxy Variables of an Unmeasured Confoundercited by 139
- 2022Semiparametric doubly robust targeted double machine learning: a reviewcited by 133
- 2014Consistency of random forestscited by 85
- 2018A confounding bridge approach for double negative control inference on causal effectscited by 36
- 2020Identifying effects of multiple treatments in the presence of unmeasured confoundingcited by 16
- 2021Data Integration in Causal Inferencecited by 11
- 2022Bagging in overparameterized learning: Risk characterization and risk monotonizationcited by 11
- 2020Assumption-lean inference for generalised linear model parameterscited by 9
- 2017Structured Latent Factor Analysis for Large-scale Data: Identifiability, Estimability, and Their Implicationscited by 8
- 2015Confounder Adjustment in Multiple Hypothesis Testingcited by 6
- 2023Is your data alignable? Principled and interpretable alignability testing and integration of single-cell datacited by 6
- 2023Extrapolated cross-validation for randomized ensemblescited by 4
10 more
- 2024Causal Inference with High-dimensional Discrete Covariatescited by 4
- 2020The Promises of Parallel Outcomescited by 3
- 2022Inference in High-dimensional Multivariate Response Regression with Hidden Variablescited by 3
- 2023Simultaneous inference for generalized linear models with unmeasured confounderscited by 3
- 2018Assumption Lean Regressioncited by 2
- 2019Doubly Robust Inference when Combining Probability and Non-probability Samples with High-dimensional Datacited by 2
- 2020The Frisch--Waugh--Lovell Theorem for Standard Errorscited by 2
- 2024Augmented Doubly Robust Post-Imputation Inference for Proteomic Datacited by 2
- 2024Substitute adjustment via recovery of latent variablescited by 1
- 2024Causal Inference for Genomic Data with Multiple Heterogeneous Outcomescited by 1
Built on by
Not cited yet.