Elucidating causal relationships between the components within complex networks has impact in commercial and research applications. For example, the ability to understand causal relationships of the components that determine stock prices can predict how the stock market will behave. In biomedical research, knowing the relationships between the genes within complex biological pathways can lead to fundamental understanding of the mechanisms of diseases such as cancer and Alzheimer’s disease and generate in-silico predictions on phenotypes which facilitates drug discovery. The current methods to analyze complex networks require the data of each component to be similar in type and can establish correlations but cannot resolve the causal relationships accurately.
Professors Eric Schadt and Rui Chang of the Icahn School of Medicine of Mount Sinai have developed a unique algorithmic approach to infer causal relationships in complex networks. This method was validated by first generating a synthetic dataset of gene expression data from a real metabolic network that is well characterized. The algorithm was then applied to this synthetic dataset and they successfully recapitulate the causal structure and recover unique features of the network which current methods cannot do. Furthermore, the inventors were successful in using the same algorithm to determine the causal relationship between the stock price of an oil company and a retail store. Using historical data of the stock prices, they determined how the price of the oil company’s stock impacted the stock price of the retail store and not vice versa. This result was consistent with published results using a different approach.
Current Development Status
- Validation in biological networks and pathways
- Validated in a financial model to determine the causal relationship between two different company stocks
Applications
- Accurately determine the causal relationships and the causal direction between components in complex networks
- Applications in a broad range of industries where large datasets are analyzed (biomedical, financial, telecommunications, etc…)
Advantages
- Ability to determine causal relationships in complex networks without the limitations of current state of the art methods
- No limitations on the nature of the data (can use continuous as oppose to discrete data)
- No limitations on the complexity and nature of the network (ie the graphs that describes the network can be of any type, such as can inference cycles (cyclic causality between nodes) in the graph.
- Ability to distinguish causality that are undistinguishable by state-of-the-arts method, e.g. can infer causality in equivalence classes of Bayesian-network structures
- Can work with static snapshot data and/or time-series observations
- Can generate future in-silico predictions on levels of the nodes in the network given perturbation
Publications
- Chang, R; Karr, J; Schadt, E. Causal Inference in Biological Networks with Integrated Belief Propagation. https://psb.stanford.edu/psb-online/proceedings/psb15/chang.pdf
- Schadt, E. et al An Integrative Genetic Approach to Infer Causal Associations Between Gene Expression and Disease. Nature Genetics, 37, 710, 2005
Patent Status
- US Provisional Application 62/046,670 filed September 5, 2014
- Status: Pending
Contact
William Chiang, PhD
Business Development Director
Mount Sinai Innovation Partners | Icahn School of Medicine at Mount Sinai
Phone: (609) 575-7033