An HMM approach to identify components that influence phenotypes

Mathematical models of biological systems have traditionally described processes that occur at the gene, enzyme, and metabolic pathway levels. We are often unable to relate these low-level models to higher-level biological phenomena. For example, models that describe the photosynthetic pathway do no...

Full description

Saved in:
Bibliographic Details
Main Authors: de Luis Balaguer, Maria A., Williams, Cranos M.
Format: Conference Proceeding
Language:English
Subjects:
Biochemistry
Biological system modeling
Biological systems
Hidden Markov models
Mathematical model
Trajectory
Vectors
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Mathematical models of biological systems have traditionally described processes that occur at the gene, enzyme, and metabolic pathway levels. We are often unable to relate these low-level models to higher-level biological phenomena. For example, models that describe the photosynthetic pathway do not explicitly relate the components to changes in photosynthetic rate. Thus, there is a need for developing methods that can link low levels of biological organization and higher-level phenotypes. We present an approach to solve this bottleneck that combines 1) a decomposition algorithm, 2) machine learning tools and 3) sensitivity analysis. With this approach, we quantified the influence of key components and functional modules on specific phenotypes, such as the photosynthetic pathway in C3 plants. Our algorithm was able to predict the relationship between specific components and the photosynthetic rate, where these results were consistent with previous experimental data. With these results we demonstrate that computational methods can be used to identify key modules and/or components that influence a measurable output of a biological system.
ISSN:1062-922X
2577-1655
DOI:10.1109/SMC.2014.6974094