MIST : MIcro Simulation Tool for Chronic Disease Modeling

MIST stands for MIcro Simulation Tool. It is free software that supports modeling and Monte-Carlo simulation. This means that simulation is conducted by generating random numbers that determine results. The word micro indicates that simulation is conducted at the micro individual level, one individual by another, rather than at the macro population level. It was initially designed to support chronic disease modeling, yet it is general enough to support other applications.

MIST includes a Domain Specific Language (DSL) aimed at creating Monte Carlo Simulations. MIST checks the model, runs it, collects results, and displays them.

MIST is capable of running the simulation in several environments, and specifically High Performance Computing (HPC) environments. Such environments include powerful multi-core computers, clusters of computers, and the Amazon Elastic Compute Cloud (EC2) where computing time can be rented. Harnessing such resources in the past was restricted to large organizations at great cost, yet the persistent lowering of computing prices make such resources available even to the single modeler. The modern modeler should delegate repetitive tasks to the computer. This can reduce uncertainty associated with Monte Carlo simulations. Simulations supported by computing power can test many more hypotheses and variations than what was possible in the past. For disease models hypotheses also include unknown information from many more initial simulation conditions defined by base populations.

In fact base populations in MIST can be defined by tabular data if the modeler has direct access to individual population data. However, MIST also allows generating mock base populations from publicly available summary data. For instance a modeler can enter the information available in the first table in a clinical trial publication into MIST and generate a set of imaginary individuals that their summary statistics matches the published clinical trial population statistics.

This capability was extended by using the Inspyred library [1] that uses evolutionary computation. This recent addition helps eliminate errors from the simulation and allows better modeling of clinical trial populations with inclusion and exclusion criteria that skew population statistics. This new modeling capability has the potential of providing help in designing the population statistics in table 1 before the clinical trial begins [2].

The Reference Model for disease progression uses MIST as the main simulation engine. Recent information about this combination is available in the video in [3].

MIST is free software and its source code and documentation can be downloaded from [4].

Further information on MIST can be found in [5].


[1] inspyred: Bio-inspired Algorithms in Python. Documentation Software Repository

[2] J. Barhak, A. Garrett, Population Generation from Statistics Using Genetic Algorithms with MIST + INSPYRED. MODSIM World 2014, April 15 - 17, Hampton Roads Convention Center in Hampton, VA. Paper Presentation

[3] J. Barhak, The Reference Model for Disease Progression uses MIST to find data fitness. PyData Silicon Valley 2014 held at Facebook Headquarters: Presentation Video

[4] MIST on Github Software Repository

[5] J. Barhak, The Reference Model for Disease Progression. SciPy 2012, Austin Tx, 18-19 July 2012. Paper Poster Video

FrontPage (last edited 2016-06-07 07:21:24 by jbarhak)