Dataset Information

Causal Discovery from Discrete Data using Hidden Compact Representation.

ABSTRACT: Causal discovery from a set of observations is one of the fundamental problems across several disciplines. For continuous variables, recently a number of causal discovery methods have demonstrated their effectiveness in distinguishing the cause from effect by exploring certain properties of the conditional distribution, but causal discovery on categorical data still remains to be a challenging problem, because it is generally not easy to find a compact description of the causal mechanism for the true causal direction. In this paper we make an attempt to find a way to solve this problem by assuming a two-stage causal process: the first stage maps the cause to a hidden variable of a lower cardinality, and the second stage generates the effect from the hidden representation. In this way, the causal mechanism admits a simple yet compact representation. We show that under this model, the causal direction is identifiable under some weak conditions on the true causal mechanism. We also provide an effective solution to recover the above hidden compact representation within the likelihood framework. Empirical studies verify the effectiveness of the proposed approach on both synthetic and real-world data.

SUBMITTER: Cai R

PROVIDER: S-EPMC6448794 | biostudies-literature | 2018 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Causal Discovery from Discrete Data using Hidden Compact Representation.

Cai Ruichu R Qiao Jie J Zhang Kun K Zhang Zhenjie Z Hao Zhifeng Z

Advances in neural information processing systems 20181201

Causal discovery from a set of observations is one of the fundamental problems across several disciplines. For continuous variables, recently a number of causal discovery methods have demonstrated their effectiveness in distinguishing the cause from effect by exploring certain properties of the conditional distribution, but causal discovery on categorical data still remains to be a challenging problem, because it is generally not easy to find a compact description of the causal mechanism for the ...[more]

PMID: 30956519

Similar Datasets

Project description:BackgroundProtein phosphorylation networks play an important role in cell signaling. In these networks, phosphorylation of a protein kinase usually leads to its activation, which in turn will phosphorylate its downstream target proteins. A phosphorylation network is essentially a causal network, which can be learned by causal inference algorithms. Prior efforts have applied such algorithms to data measuring protein phosphorylation levels, assuming that the phosphorylation levels represent protein activity states. However, the phosphorylation status of a kinase does not always reflect its activity state, because interventions such as inhibitors or mutations can directly affect its activity state without changing its phosphorylation status. Thus, when cellular systems are subjected to extensive perturbations, the statistical relationships between phosphorylation states of proteins may be disrupted, making it difficult to reconstruct the true protein phosphorylation network. Here, we describe a novel framework to address this challenge.ResultsWe have developed a causal discovery framework that explicitly represents the activity state of each protein kinase as an unmeasured variable and developed a novel algorithm called "InferA" to infer the protein activity states, which allows us to incorporate the protein phosphorylation level, pharmacological interventions and prior knowledge. We applied our framework to simulated datasets and to a real-world dataset. The simulation experiments demonstrated that explicit representation of activity states of protein kinases allows one to effectively represent the impact of interventions and thus enabled our framework to accurately recover the ground-truth causal network. Results from the real-world dataset showed that the explicit representation of protein activity states allowed an effective and data-driven integration of the prior knowledge by InferA, which further leads to the recovery of a phosphorylation network that is more consistent with experiment results.ConclusionsExplicit representation of the protein activity states by our novel framework significantly enhances causal discovery of protein phosphorylation networks.

Project description:BACKGROUND:This study demonstrates the use of reduced-representation genotyping to provide preliminary identifications for thermophilic bacterial isolates. The approach combines restriction enzyme digestion and PCR with next-generation sequencing to provide thousands of short-read sequences from across the bacterial genomes. Isolates were obtained from compost, hot water systems, and artesian bores of the Great Artesian Basin. Genomic DNA was double-digested with two combinations of restriction enzymes followed by PCR amplification, using a commercial provider of DArTseq™, Diversity Arrays Technology Pty Ltd. (Canberra, Australia). The resulting fragments which formed a reduced-representation of approximately 2.3% of the genome were sequenced. The sequence tags obtained were aligned against all available RefSeq bacterial genome assemblies by BLASTn to identify the nearest reference genome. RESULTS:Based on the preliminary identifications, a total of 99 bacterial isolates were identified to species level, from which 8 isolates were selected for whole-genome sequencing to assess the identification results. Novel species and strains were discovered within this set of isolates. The preliminary identifications obtained by reduced-representation genotyping, as well as identifications obtained by BLASTn alignment of the 16S rRNA gene sequence, were compared with those derived from the whole-genome sequence data, using the same RefSeq sequence database for the three methods. Identifications obtained with reduced-representation sequencing agreed with the identifications provided by whole-genome sequencing in 100% of cases. The identifications produced by BLASTn alignment of 16S rRNA gene sequence to the same database differed from those provided by whole-genome sequencing in 37.5% of cases, and produced ambiguous identifications in 50% of cases. CONCLUSIONS:Previously, this method has been successfully demonstrated for use in bacterial identification for medical microbiology. This study demonstrates the first successful use of DArTseq™ for preliminary identification of thermophilic bacterial isolates, providing results in complete agreement with those obtained from whole-genome sequencing of the same isolates. The growing database of bacterial genome sequences provides an excellent resource for alignment of reduced-representation sequence data for identification purposes, and as the available sequenced genomes continue to grow, the technique will become more effective.

Dataset Information

Causal Discovery from Discrete Data using Hidden Compact Representation.

Publications

Causal Discovery from Discrete Data using Hidden Compact Representation.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets