miniTUBA

Category Intelligent Software>Bayesian Network Systems/Tools

Abstract miniTUBA stands for the Medical Inference by Network Integration of Temporal Data using Bayesian Analysis. It is a web-based dynamic Bayesian Network analysis system used to simulate biomedical networks using temporal data.

The results from these analyses can then be utilized to identify causal or potentially causal relationships between data elements.

The primary users of miniTUBA are bioinformaticians, biomedical researchers, and clinicians. Users will be able to continuously update their data and refine their results. miniTUBA will offer prediction and intervention suggestions based on an automatic learning process pipeline using all data provided.

The Dynamic Bayesian Networks (DBN) analysis pipeline developed by miniTUBA is widely applicable (useful well beyond just biological data) and can be useful for generating networks using Not only clinically- relevant data.

Bayesian Network analysis --

Dynamic Bayesian Networks (DBN) represents an advanced probabilistic modeling method for identifying causal or apparently causal patterns in temporal clinical and biomedical research data. Using Dynamic Bayesian networks, an investigator can identify both sets of significantly regulated parameters (clinical and/or molecular) and also the interrelationships between these variables.

This approach is particularly useful when there are a large number of variables and their interactions are complicated. A key advantage of Bayesian networks for identifying apparently causal relationships between multiple variables is that Bayesian networks are relatively agnostic to the complexity of the relationships predicted.

For example, a Bayesian network can model linear, nonlinear, or more complicated multi-state relationships equally well. Additionally, Bayesian networks are able to create chains of causation, suggesting a sequence of events required to produce a particular outcome.

Such chains of causation are difficult to extract using other machine learning techniques.

In addition, by modifying the structural priors (e.g. forcing or forbidding certain connections) a Dynamic Bayesian Network can also include 'expert knowledge' to constrain the results.

miniTUBA data --

miniTUBA currently contains both synthetic data and real biomedical research data.

Sandbox Demo --

miniTUBA offers a Sandbox Demo system, which allows you to login without registration and test all the features in miniTUBA using the miniTUBA demo account. Over twenty (20) projects currently exist in the Sandbox Demo system.

Note: Since every one has access to this site, you may Not want to use your own research data in the Sandbox Demo system and with the demo account, a user can only use up to 10 minutes of computation time and up to two (2) computer servers in the miniTUBA multi-server system.

Research Projects --

miniTUBA currently host ten (10) 'public' research projects and a number of private projects.

You can learn more about miniTUBA from the following Bioinformatics paper:

miniTUBA: medical inference by network integration of temporal data using Bayesian analysis -- Bioinformatics 2007 23(18):2423-2432; doi: 10.1093/bioinformatics/btm372, PMID: 17644819

System Requirements

Contact manufacturer.

Manufacturer

Manufacturer Web Site miniTUBA

Price Contact manufacturer.

G6G Abstract Number 20205

G6G Manufacturer Number 102855