#### DMCA

## Constraintbased causal discovery from multiple interventions over overlapping variable sets. arXiv:1403.2150 (2014)

Citations: | 2 - 0 self |

### Citations

802 | Statistical significance for genomewide studies - Storey, Tibshirani - 2003 |

672 | An extensible SAT-solver, in - Eén, Sörensson |

296 | Causal protein-signaling networks derived from multiparameter single-cell data - Sachs, Perez, et al. - 2005 |

270 |
Equivalence and synthesis of causal models
- Verma, Pearl
- 1990
(Show Context)
Citation Context ...t, that the indistinguishability of M1 and M2 refers to m-separation only; the absence of a direct causal edge between A and D could be detected using other types of tests, like the Verma constraint (=-=Verma and Pearl, 2003-=-). While we cannot predict the effect of manipulations on a MAG M, given a data set measuring variables O when variables in I ⊂ O are manipulated, we can obtain (assuming an oracle of conditional inde... |

225 | Cytoscape 2.8: New features for data integration and network visualization - Smoot, Ono, et al. - 2011 |

163 | H (2000) Heavy-tailed phenomena in satisfiability and constraint satisfaction problems
- CP, Selman, et al.
(Show Context)
Citation Context ... SBCSD since it did not complete for most problems). Thus, performance highly depends on the input structure. Such heavy-tailed distributions are well-noted in the constraint satisfaction literature (=-=Gomes et al., 2000-=-). We also note the fact that COmbINE seems to depend more on the sparsity and less on the number of variables, while SBCSD’s time increases monotonically with the number of variables. Based on these ... |

154 |
Learning gaussian networks
- Geiger, Heckerman
- 1994
(Show Context)
Citation Context ... for FCI is limited to 3 in all experiments in this section for a fair comparison. For a (non) collider X Y Z, we score all networks over X, Y and Z. We use the BGE metric for gaussian distributions (=-=Geiger and Heckerman, 1994-=-) as implemented in the BDAGL package Eaton and Murphy (2007b) to calculate the likelihoods of the DAGs. This metric is score equivalent, so we pre-computed representatives of the Markov equivalent ne... |

126 | Ancestral graph Markov models - Richardson, Spirtes - 2002 |

79 | Causal discovery from a mixture of experimental and observational data - Cooper, Yoo - 1999 |

75 | Active learning for structure in Bayesian networks. - Tong, Koller - 2001 |

60 | Active learning of causal bayes net structure - Murphy - 2001 |

30 | On the identification of causal effects
- Tian, Pearl
- 2003
(Show Context)
Citation Context ...ich all fitting models agree, as well as the ones for which conflicting explanations are plausible. As our formalism of choice for causal modeling, we employ Semi-Markov Causal Models (SMCMs). SMCMs (=-=Tian and Pearl, 2003-=-) are extensions of Causal Bayesian Networks (CBNs) that can account for latent confounding variables, but do not admit feedback cycles. Internally, the algorithm also makes heavy use of the theory an... |

28 | p38 kinase-dependent MAPKAPK-2 activation functions as 3-phosphoinositide-dependent kinase-2 for Akt in human neutrophils - Rane - 2001 |

25 | Single-cell mass cytometry of differential immune and drug responses across a human hematopoietic continuum. - Bendall - 2011 |

25 | On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias. - Zhang - 2008 |

20 | A statistical problem for inference to regulatory structure from associations of gene expression measurements with microarrays
- Chu, Glymour, et al.
(Show Context)
Citation Context ...be an suitable test-bed for causal discovery methods: The proteins are measured in single cells instead of representing tissue averages, the latter being known to be problematic for causal discovery (=-=Chu et al., 2003-=-), and the samples range in thousands. However, the mass cytometer can measure only up to 34 variables, which may be too low a number to measure all the variables involved in a signaling pathway. More... |

18 | Learning high-dimensional directed acyclic graphs with latent and selection variables - Colombo, Maathuis, et al. |

17 | A Polynomial Time Algorithm For Determining DAG Equivalence - Spirtes, Richardson - 1996 |

15 | 2008]: ‘Integrating Locally Learned Causal Structures with Overlapping Variables - Tillman, Danks, et al. |

14 | Interventions and Causal Inference”. - Eberhardt, Scheines - 2007 |

11 | Multiplexed mass cytometry profiling of cellular states perturbed by small-molecule regulators. - Bodenmiller, ER, et al. - 2012 |

11 | A Bayesian approach to constraint based causal inference
- Claassen, Heskes
(Show Context)
Citation Context ...g to Bayesian probability estimates of the literals in F as presented in Claassen and Heskes (2012). The same greedy strategy for satisfying constraints in order is employed. Briefly, the authors of (=-=Claassen and Heskes, 2012-=-) propose a method for calculating Bayesian probabilities for any feature of a causal graph (e.g. adjacency, m-connection, causal ancestry). To estimate the probability of a feature, for a given data ... |

11 | Marginal log-linear Parameters for Graphical Markov Models. - Evans, Richardson - 2013 |

10 | Causal Inference and Reasoning in Causally Insufficient Systems
- Zhang
- 2006
(Show Context)
Citation Context ...anipulated variable should not directly affect any variable other than its direct target, and more importantly, local mechanisms for other variables should remain the same as before the intervention (=-=Zhang, 2006-=-). Thus, the intervention is merely a local surgery with respect to causal mechanisms. These assumptions may seem a bit restricting, but this type of experiment is fairly common in several modern fiel... |

9 | Learning linear cyclic causal models with latent variables - Hyttinen, Eberhardt, et al. |

8 | Characterization and greedy learning of interventional Markov equivalence classes of directed acyclic graphs - Hauser, Bühlmann - 2012 |

7 | Maximum likelihood fitting of acyclic directed mixed graphs to binary data - Evans, Richardson - 2010 |

6 | Discovering cyclic causal models with latent variables: A general sat-based procedure - Hyttinen, Hoyer, et al. - 2013 |

6 | Learning equivalence classes of acyclic models with latent and selection variables from multiple datasets with overlapping variables - Tillman, Spirtes - 2011 |

5 | Causal discovery of linear cyclic models from multiple experimental data sets with overlapping variables - Hyttinen, Eberhardt, et al. |

4 |
Improved exact solver for the weighted max-sat problem
- Kuegel
- 2010
(Show Context)
Citation Context ...inguish between hard and soft constraints. To maximize the number of literals satisfied, while ensuring all hard-constraints are satisfied we resorted to the following technique: we use the akmaxsat (=-=Kuegel, 2010-=-) weighted max SAT solver that tries to maximize the sum of the weights of the satisfied clauses. Each literal is assigned a weight of 1, and each hard-constraint is assigned a weight equal to the sum... |

4 | Towards integrative causal analysis of heterogeneous data sets and studies - Tsamardinos, Triantafillou, et al. - 2012 |

3 | Learning causal network structure from multiple (in)dependence models,” - Claassen, Heskes - 2010 |

3 | Belief net structure learning from uncertain interventions - Eaton, Murphy |

3 | Learning semi-markovian causal models using experiments - Meganck, Maes, et al. - 2006 |

2 |
Tools and algorithms for causally interpreting directed edges in maximal ancestral graphs
- Borboudakis, Triantafillou, et al.
- 2012
(Show Context)
Citation Context ...and which are there because of a (non-trivial) primitive inducing path. Note though that, there exist sound and complete algorithms that identify all edges for which such a determination is possible (=-=Borboudakis et al., 2012-=-). In addition, we later show that co-examining manipulated distributions can indicate that some edges stand for indirect causality (or indirect confounding). 11 3.4 Manipulations under causal insuffi... |

2 | Nested markov properties for acyclic directed mixed graphs - Richardson, Robins, et al. - 2012 |

2 | Calibration of ρ values for testing precise null hypotheses - Sellke, Bayarri, et al. |

2 | IG Tollis. Learning causal structure from overlapping variable sets - Triantafillou, Tsamardinos |

1 | Bdagl: Bayesian dag learning. http://www.cs.ubc.ca/~murphyk/ Software/BDAGL - Eaton, Murphy |

1 | Sparse nested markov models with log-linear parameters - Shpitser, Evans, et al. - 2013 |