Some algae produce in excess of 40% of their dry mass as fatty acids that can be synthesized into fuel. This presentation concerns identifying superior environmental and media conditions for the increased production of these fats. Parameters that influence the biosynthesis of lipids include light intensity, nutrient concentration, and temperature. In order to assess the quantity of fatty acids per cell, the lipids must be converted into a form better suited for gas chromatography, known as fatty acid methyl esters (FAMEs). To transesterify the raw oil an algal sample is heated for one hour with boron trifluoride and methanol producing the methyl esters as the desired products. A subsequent sub-microscale liquid-liquid extraction is performed to separate the esters from the parent algal residue. In order to quantify the amount each fatty acid present (which may have varying chain lengths and degrees of unsaturation), the mixture of esters is run through a GC/MS against an internal standard. Using this method, it has been determined that an increase in concentration of sodium acetate in the growth media positively influences the amount of fatty acid per cell in the model organism. Results of this experiment have shown a thirty percent increase over control group fatty acid production while changes to other parameters have not shown corresponding increases. These experiments show great promise, as sodium acetate is both inexpensive and can be derived from biological matter. Should these results be confirmed, productivity of fatty acid synthesis can be increased at minimal cost using nontoxic media additives with the model organism. Tests need to be conducted in other strains of algae to determine if a similar effect is observed.

When studying cellular networks computationally, visualizing the networks on screen can be a useful aid to the researcher. Currently, there are a handful of programs that exist for this purpose. They are all desktop applications that need to be installed on the computer before they can be used. One of the most common programs found on computers is the Internet browser and its prevalence makes it an interesting backbone on which to build a visualization tool. When building an application in a browser, the installation process can be avoided and differences in varying operating systems can be disregarded. A standard browser plug-in is Adobe’s Flash. The purpose of this study is to develop and evaluate Flash as the platform on which to build a web based pathway modeling tool. Using ActionScript (the programming language used to create Flash programs), I assessed the ability of Flash to create, edit and save representations of cellular networks in a simple and efficient manner. The application can form a pathway by listening to user events or by reading an input file. Once created, these pathways can be saved in three different formats: JPEG image, text file, and Systems Biology Markup Language. The ActionScript language fosters rapid graphical user interface development and offers many unique features with its library of classes. ActionScript has many built-in features; in particular, the process of creating visible objects and listening to user events has been streamlined. CS4 Professional, the environment used to write ActionScript code, is an integral part of the programming experience. It enables programmers to easily create readable code, pinpoint errors and publish their work in websites. In this study, Flash was found to be an exceptional platform for making this type of modeling tool.

Hematopoiesis is the formation, development and specialization of the various blood cells from the hematopoietic stem cells (HSC). Each HSC can replicate, die, or differentiate into progenitor cells. Whereas the behavior of progenitor cells both in vivo and in vitro has been studied extensively, the behavior of HSCs is hardly known since the exact structure and interrelationships of HSC compartment are unclear and cannot be recreated in vitro. We develop a visualization tool for a two-compartment hidden stochastic process used to model the kinetics of hematopoietic stem cells in vivo. This Java-based program visualizes the stochastic model over time. It not only provides an easy to use graphical user interface (GUI), but also offers interactive means to modify the model parameters and control the simulation process. It produces both graphical outputs and summary statistics. Meanwhile, the portability of Java code removes the platform restriction on the program. In addition, we have developed a comprehensive user manual containing instructions for using the program along with background information on hematopoiesis, stem cell theory, stochastic processes, and Markov chain Monte Carlo methods so that individuals without intensive knowledge in hematology or statistics could take advantage of the simulator. This program will be an effective visualization tool for modeling both normal and abnormal hematopoiesis, aid researchers to gain intuitive understanding of the processes, and offer the opportunities for explorations of virtual experiments.

One motivation for examining medial graphs arises from the study of the electrical conductivity inverse problem. A typical presentation of the inverse problem is as follows: Suppose we are given a network of conductive nodes that are connected by resistors. Designate some of these nodes as being on the boundary of the network. The inverse problem is to use measurements of voltage and current on these boundary nodes to determine the conductivities of the resistors inside the network. Now suppose that the graph of nodes and resistors underlying this network is unknown. Is it possible to use our boundary measurements to recover this graph of this network and in turn recover its conductivities? If we restrict our attention to certain classes of graphs, then the answer is affirmative. The recovery of the graph of a network goes by way of first constructing its medial graph; once we have a medial graph, we can reconstruct its corresponding network with relative ease. We present an algorithm for using boundary voltage and current information to determine this medial graph. Our approach relies heavily on matrix computations and has features to improve its robustness to measurement and floating point error. We then demonstrate an algorithm and Sage implementation for displaying this graph that avoids some of the pitfalls of automated graph generation. At the heart of our method, we parametrize symmetric parabolic arcs so that they avoid undesirable intersections with each other. It is our hope to use these computational tools to improve our understanding of medial graphs and electrical networks, especially when it comes to classifying the possible combinations of boundary voltage and current measurements.