The focus of this study is on predicting the subcellular localization of a protein. Subcellular localization information is important for protein function annotation which is a fundamental problem in computational biology. For this problem, a classification system is built that has two main parts: a predictor that is based on a feature mapping technique to extract biologically meaningful information from protein sequences and a client/server architecture for searching and predicting subcellular localizations. In the first part of the thesis, we describe a feature mapping technique based on frequent patterns. In the feature mapping technique we describe, frequent patterns in a protein sequence dataset were identified using a search technique based on a priori property and the distribution of these patterns over a new sample is used as a feature vector for classification. The effect of a number of feature selection methods on the classification performance is investigated and the best one is applied. The method is assessed on the subcellular localization prediction problem with 4 compartments (Endoplasmic reticulum (ER) targeted, cytosolic, mitochondrial, and nuclear) and the dataset is the same used in P2SL. Our method improved the overall accuracy to 91.71% which was originally 81.96% by P2SL. In the second part of the thesis, a client/server architecture is designed and implemented based on Simple Object Access Protocol (SOAP) technology which provides a user-friendly interface for accessing the protein subcellular localization predictions. Client part is in fact a Cytoscape plug-in that is used for functional enrichment of biological networks. Instead of the individual use of subcellular localization information, this plug-in lets biologists to analyze a set of genes/proteins under system view.

Indiana University-Purdue University Indianapolis (IUPUI) / Metabolic flux analyses and their more comprehensive forms, genome-scale metabolic networks (GSMNs), have gained tremendous attention in industrial and medical research. Saccharomyces cerevisiae (S. cerevisiae) is one of the organisms that has had its GSMN subjected to multiple frequent updates. The objective of this study is to develop a visualization tool for the GSMN of S. cerevisiae for educational and research purposes. This visualization tool is called the Master Metabolic Map of Saccharomyces cerevisiae (MMMSC). In this study, a metabolic database of S. cerevisiae developed by us was transferred to Cytoscape, a useful and efficient bioinformatics software platform for visualizing molecular networks. After the MMMSC was created, nodes, representing metabolites and enzymes, and edges, representing the chemical reactions that connect the nodes, were curated manually to develop a metabolic visualization map of the whole metabolic system of S. cerevisiae (Figure 4). In the discussion, examples are provided regarding possible applications of MMMSC to predict possible effects of the manipulation of the S. cerevisiae metabolome for industrial and medical purposes. Ultimately, it is concluded that further work is needed to complete the metabolic database of S. cerevisiae and the related MMMSC. In future studies, these tools may be integrated with other omics and other approaches, especially the directed-evolution approach, to increase cost and time efficiency of future research and to find solutions for complex and, thus far, poorly managed environmental and health problems.

Background Cancer is one of the leading causes of death in Sweden. In order to develop better treatments against cancer we need to better understand it. One area of special interest is cancer metabolism and the metabolic fluxes. As these fluxes cannot be directly measured, modeling is required to determine them. Due to the complexity of cell metabolism, some limitations in the metabolism model are required. As the TCA-cycle (TriCarboxylic Acid cycle) is one of the most important parts of cell metabolism, it was chosen as a starting point. The primary goal of this project has been to evaluate the previously constructed TCA-cycle model. The first step of the evaluation was to determine the CI (Confidence Interval) of the model parameters, to determine the parameters’ identifiability. The second step was to validate the model to see if the model could predict data for which the model had not been trained for. The last step of the evaluation was to determine the uncertainty of the model simulation. Method The TCA-cycle model was created using Isotopicaly labeled data and EMUs (ElementaryMetabolic Units) in OpenFlux, an open source toolbox. The CIs of the TCA-cycle model parameters were determined using both OpenFlux’s inbuilt functionality for it as well as using amethod called PL (Profile Likelihood). The model validation was done using a leave one out method. In conjunction with using the leave on out method, a method called PPL (Prediction Profile Likelihood) was used to determine the CIs of the TCA-cycle model simulation. Results and Discussion Using PL to determine CIs had mixed success. The failures of PL are most likely caused by poor choice of settings. However, in the cases in which PL succeeded it gave comparable results to those of OpenFLux. However, the settings in OpenFlux are important, and the wrong settings can severely underestimate the confidence intervals. The confidence intervals from OpenFlux suggests that approximately 30% of the model parameters are identifiable. Results from the validation says that the model is able to predict certain parts of the data for which it has not been trained. The results from the PPL yields a small confidence interval of the simulation. These two results regarding the model simulation suggests that even though the identifiability of the parameters could be better, that the model structure as a whole is sound. Conclusion The majority of the model parameters in the TCA-cycle model are not identifiable, which is something future studies needs to address. However, the model is able to to predict data for which it has not been trained and the model has low simulation uncertainty.