Permutation-based Causal Inference Algorithms with Interventions

By Yuhao Wang, Liam Solus, Karren D. Yang, Caroline Uhler. 2017

169 citations recorded by the source corpus.

Learning directed acyclic graphs using both observational and interventional data is now a fundamentally important problem due to recent technological developments in genomics that generate such single-cell gene expression data at a very large scale. In order to utilize this data for learning gene regulatory networks, efficient and reliable causal inference algorithms are needed that can make use of both observational and interventional data. In this paper, we present two algorithms of this type and prove that both are consistent under the faithfulness assumption. These algorithms are interventional adaptations of the Greedy SP algorithm and are the first algorithms using both observational and interventional data with consistency guarantees. Moreover, these algorithms have the advantage that they are nonparametric, which makes them useful also for analyzing non-Gaussian data. In this paper, we present these two algorithms and their consistency guarantees, and we analyze their performance on simulated data, protein signaling data, and single-cell gene expression data.

Subfields: Artificial Intelligence, Molecular Biology, Causal Inference & Discovery

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Molecular Biology

In Causal Inference & Discovery

Citations in the corpus

Differentiable Causal Discovery from Interventional Data (2020, methodology): “Among these, IGSP [47, 48] is a method that optimizes a score based on conditional independence tests.”
Differentiable Causal Discovery Under Latent Interventions (2022, unknown): “HD tp fn fp rev F 1 score GIES (Hauser and B¨uhlmann, 2012) 38 10 0 41 7 0.33 CAM (B¨uhlmann et al., 2014) 35 12 1 30 4 0.51 IGSP (Wang et al., 2017) 18 4 6 5 7 0.42 DCDI-G (Brouillard et al., 2020) 36 6 2 25 9 0.31 DCDI-DSF (Brouillard et al., 2020) 2020), with the exception of FCI.”
A Survey of Learning Causality with Data: Problems and Methods (2018, background): “[142] proposed algorithms to learn causal relationships when a mixture of interventional and observational data is given, which are non-parametric and handle non-Gaussian data well.”
Characterizing and Learning Equivalence Classes of Causal DAGs under Interventions (2018, methodology): “GIES is an extension of GES that incorporates interventional data into the score function it uses to search over the space of DAGs (Hauser & B ¨ uhlmann, 2012), but it is in general not consistent (Wang et al., 2017).”
A Survey of Learning Causality with Data (2018, background): “raphs exhaustively. Therefore, heuristics such as GES [26] and its extension, Fast GES (FGES) [123] are proposed to reach a locally optimal solution. When it comes to interventional data, Wang et al. [157] propose algorithms to learn causal relations when a mixture of interventional and observationa”
Predicting transcriptional outcomes of novel multigene perturbations with GEARS (2023, unknown): “Further information on research design is available in the Nature Port-folio Reporting Summary linked to this article.”
Representation Learning for Treatment Effect Estimation from Observational Data (2018, methodology): “Owing to the easy access of observed data, observational studies, such as the potential outcome framework [27] and causal graphical models [26, 35], have been widely applied in various domains [15, 38, 12].”
Characterization and Learning of Causal Graphs with Latent Variables from Soft Interventions (2019, background): “Related Work: Learning causal graphs from a combination of observational and interventional data has been studied in the literature [3, 11, 7, 20, 8, 12, 23].”
Graphical criteria for efficient total effect estimation via adjustment in causal linear models (2019, unknown): “…f r o m h tt p s :/ / a c ade m i c . oup . c o m / j r sss b / a r t i c l e / 84 / 2 / 579 / 7056108 b y R o y a l L i b r a r y , C ope nhagen U n i v e r s i t y L i b r a r y u s e r on 13 A p r il 2023 Bühlmann, 2012; Hoyer et al., 2008; Meek, 1995; Scheines et al., 1998; Wang et al., 2017).”
Permutation-Based Causal Structure Learning with Unknown Intervention Targets (2019, methodology): “The proposed algorithm is an interventional adaptation of the greedy sparsest permutation (GSP) algo-rithm (Solus et al., 2017) that was proposed for causal inference in the purely observational setting.”