PB2020: Introduction to Modeling

Welcome to the workshop! You can access the workshop here: (TBD)

Do you want feedback on a modeling problem? Post in the Plant Bio 2020 forum: https://www.initmathbio.com/imbforums/

You need to be registered and logged in to see the forum. You’ll get an email notification when we respond to your question.

Contents

Pre-Workshop Survey

Workshop Agenda

Vocabulary

Worksheets (in order)

More Information

Pre-Workshop Survey

Please fill out so we understand you better and have your email!

Workshop Agenda

The workshop will be from 2-4 PM EST.

  1. Introduction (10 min)
  2.  The toy problem (10 min)
  3.  Making a concept diagram (10 min)
  4.  Presentations:
    1.  Dr Bucksch
    2.  Scott Oswald
    3.  10 minute break
    4. Dr Apka
    5.  Dr Dale
  5.  Wrap-up (5 min)
  6. Q&A (10 min)

Vocabulary

Deterministic: Assumes most things behave like the average thing

Stochastic: The process includes randomness or variability

Spatial: With respect to space

Dynamic: With respect to time

Stable state: State where things are constant. Same concept as equilibrium and steady state

Discrete: Integer number or time (eg, 1,2,3,4)

Continuous: Number with digits after the decimal point (eg, 1.29, 4.5672)

Finite: Countable number of things, as opposed to infinite

Black Box: Include a mechanism without an identified or experimentally proven component

Ball and Stick: A method of representing a conceptual model showing relationships and mechanisms between components

Parameter: A number, can represent a biological rate constant or something

Variable: A model component

Multi-scale: A model or process that spans multiple biological scales (eg, cell to tissue)

Worksheets (in order)

Worksheet 1: Box and Stick Diagram

Worksheet 2: Presentation 1: Dr Bucksch

Worksheet 3: Presentation 2: Scott Oswald

Worksheet 4: Presentation 3: Dr Apka

Worksheet 5: Presentation 4: Dr Dale

Wrap-Up Survey

More Information

Hardy-Weinberg stochastic simulation: https://rdale1.shinyapps.io/app_sept_12/

Brownian motion simulation: http://labs.minutelabs.io/Brownian-Motion/

Python code for cellular automata: (TBA)

Python code for diffusion simulation: (TBA)

Python code for root growth simulation: https://www.dropbox.com/s/fxa5wtqv71pcadj/LSystemDemo.py?dl=0

Common types of models:

  • Ordinary differential equations (ODEs) – model that describes how things change in one dimension (eg, time)
  • Flux balance (FBA) – usually a whole-organism comprehensive model containing all reactions, genes, proteins in an organism. This makes it usually restricted to single-cell organisms. Evaluated at a single time point
  • Metabolic flux (mFVA) – similar to FBA, but a subset of the whole organism, so that eukaryotes can be modeled (including plants)
  • Morphology models – models that describe shape and form
  • Cellular automata or agent-based models – rules, whether stochastic or deterministic, guide the behavior of system components
  • Stochastic differential equations (SDEs) – model where one or more components have a variability to them in how they change over a dimension (eg, time)
  • Probabilistic or Markov models – models with strong stochastically driven process
  • Multi-scale models – various mathematical components, but contain multiple scales or time scales (things that happen quickly, like enzyme reactions, and slowly, like reproduction)
  • Partial differential equations (PDEs) – can change in more than one dimensions (eg, time and space)

What kind of questions can models test?

Questions about models/model structure, math representation, interactions/strength of interactions, questions about importance of components, certain experimental design strategies, mechanisms or relationships

What makes modeling a science, vs a “test” like a t-test or ANOVA?

  • Unique (or so) modeling representation for each biological question/system
  • Subjective 
  • High degree of inferential power but low generalizability (usually) – because its generalizable to abstract system, not biological system
  • Although you COULD black box a model, you may make really bizarre assumptions about your system. Some tools exist that get around this by having a UI so you can make connections and specify their biological importance directly
  • Mathematics and equation structure provides biological inference beyond simply testing something (eg, global behavior, phase shifts, bistability)

More vocabulary:

Global behavior

Parameter optimization

Phase shift:

Bi-stability