Machine Learning and Mapping Algorithms Applied to Proteomics Problems
Sanders, W. S. (2011). Machine Learning and Mapping Algorithms Applied to Proteomics Problems. Mississippi State University: ProQuest/UMI Dissertation Publishing.
Proteins provide evidence that a given gene is expressed, and machine learning
algorithms can be applied to various proteomics problems in order to gain information
about the underlying biology. This dissertation applies machine learning algorithms to
proteomics data in order to predict whether or not a given peptide is observable by mass
spectrometry, whether a given peptide can serve as a cell penetrating peptide, and then
utilizes the peptides observed through mass spectrometry to aid in the structural
annotation of the chicken genome. Peptides observed by mass spectrometry are used to
identify proteins, and being able to accurately predict which peptides will be seen can
allow researchers to analyze to what extent a given protein is observable. Cell
penetrating peptides can possibly be utilized to allow targeted small molecule delivery
across cellular membranes and possibly serve a role as drug delivery peptides. Peptides
and proteins identified through mass spectrometry can help refine computational gene
models and improve structural genome annotations.