Thesis (M.S.P.H.)--University of Michigan, 1941.

In the neuronal circuits of natural and artificial agents, memory is usually implemented with recurrent connections, since recurrence allows past agent state to affect the present, on-going behavior. Here, an interesting question arises in the context of evolution: how reactive agents could have evolved into cognitive ones with internalized memory? This study strives to find an answer to the question by simulating neuroevolution on artificial neural networks, with the hypothesis that internalization of external material interaction can be a plausible evolutionary path leading to a fully internalized memory system. A series of computational experiments were performed to gradually verify the above hypothesis. The first experiment demonstrated the possibility that external materials can be used as memory-aids for a memoryless reactive artificial agents in a simple 1-dimensional environment. Here, the reactive artificial agents used environmental markers as memory references to be successful in the ball-catching task that requires memory. Motivated by the result of the first experiment, an extended experiment was conducted to tackle a more complex memory problem using the same principle of external material interaction. This time, the reactive artificial agents are tasked to remember the locations of food items and the nest in a 2-dimensional environment. Such path-following behavior is a trivial foraging strategy of various lower animals such as ants and fish. The final experiment was designed to show the evolution of internal recurrence. In this experiment, I showed the evolutionary advantage of external material interaction by comparing the results from neural network topology evolution algorithms with and without the material interaction mechanism. The result confirmed that the agents with external material interaction learned to solve the memory task faster and more accurately. The results of the experiments provide insights on the possible evolutionary route to an internalized memory. The use of external material interaction can help reactive artificial agents to go beyond the functionality restricted by their simple network structure. Moreover, it allows much faster convergence with higher accuracy than the topological evolution of the artificial agents. These results suggest one plausible evolutionary path from reactive, through external material interaction, to recurrent structure.

Understanding the origin of life requires understanding the origin of translation, which in turn, requires understanding the origin of the ribosome. Ribosomes are complex structures consisting of hundreds of thousands of atoms. Here, we describe how we organized ribosomal structures and information into a broad database, RiboZones. We also describe a new visualization web app, RiboVision. RiboZones and RiboVision are productivity tools that lower the learning curve for ribosomal research. RiboZones makes the ribosome more accessible. RiboVision especially helps create beautiful publication ready figures in a fraction of the labor and time previously required. It is only through the creation of RiboZones and RiboVision through which the rest of this dissertation became feasible. We constructed a high-quality sequence alignment of ribosomal sequences for both the LSU and the SSU rRNA. Each ribosomal sequence is complete, allowing detailed, low background statistics to be computed. The sequence alignment broadly samples the tree of life according to available data. The alignment was adjusted for maximum agreement with 3D superimpositions of multiple ribosomal structures. We defined a nucleotide-level definition of the common core of the ribosome, as the RNA that is present in 95% of the sequences in our alignment. Multiple versions of the common core were created, including the universal common core, the prokaryotic common core, and domain specific common cores. The definition allows statistics to be computed for various use-cases. For example, with RiboVision visualization technology, it is possible to see which helices are optional, in which of the three domains of life, and what the minimum helical length is for each helix. We discovered that ribosomal RNA grows mostly by helix extension and helix insertion. When a helix is inserted, it minimally perturbs the underlying helix. We call this pattern ‘insertion fingerprints’. Insertion fingerprints are found throughout the common core and the eukaryotic expansion segments. Insertion fingerprints were used to divide the ribosomal RNA into units called ancestral expansion segments (AES’s). AES’s make ideal structural, functional, and evolutionary units. The AES’s are arranged into the first complete experimentally testable model of ribosomal evolution. The model can be refined over time as new information is discovered.

Based on a review of different Darwinian paradigms, particularly sociobiology, this work, both, historically and philosophically, develops a metaphysic of gene-Darwinism and process-Darwinism, and then criticises and transcends these Darwinian paradigms in order to achieve a truly evolutionary theory of evolution. Part I introduces essential aspects of current sociobiology as the original challenge to this investigation. The claim of some sociobiologists that ethics should become biologized in a gene-egoistic way, is shown to be tied to certain biological views, which ethically lead to problematic results. In part II a historical investigation into sociobiology and Darwinism in general provides us, as historical epistemology', with a deeper understanding of the structure and background of these approaches. Gene-Darwinism, which presently dominates sociobiology and is linked to Dawkins' selfish gene view of evolution, is compared to Darwin's Darwinism and the evolutionary' synthesis and becomes defined more strictly. An account of the external history of Darwinism and its subparadigms shows how cultural intellectual presuppositions, like Malthusianism or the Newtonian concept of the unchangeable laws of nature, also influenced biological theory' construction. In part III universal 'process-Darwinism' is elaborated based on the historical interaction of Darwinism with non-biological subject areas. Building blocks for this are found in psychology, the theory of science and economics. Additionally, a metaphysical argument for the universality of process- Darwinism, linked to Hume's and Popper's problem of induction, is proposed. In part IV gene-Darwinism and process-Darwinism are criticised. Gene-Darwinism—despite its merits—is challenged as being one-sided in advocating 'gene-atomism', 'germ-line reductionism' and 'process-monism'. My alternative proposals develop and try to unify different criticisms often found. In respect of gene-atomism I advocate a many-level approach, opposing the necessary radical selfishness of single genes. I develop the concept of higher-level genes, propose a concept of systemic selection, which may stabilise group properties, without relying on permanent group selection and extend the applicability of a certain group selectionist model generally to small open groups. Proposals of mine linked to the critique of germ-line reductionism are: 'exformation', phenotypes as evolutionary factors and a field theoretic understanding of causa formalis (resembling Aristotelian hylemorphism). Finally the process-monism of gene-Darwinism, process-Darwinism and, if defined strictly, Darwinism in general is criticised. 1 argue that our ontology and ethics would be improved by replacing the Newtoman-Paleyian deist metaphor of an eternal and unchangeable law of nature, which lies at tire very heart of Darwinism, by a truly evolutionary understanding of evolution where new processes may gain a certain autonomy. All this results in a view that I call 'ecological idealism', which, although still very much based on Darwinism, clearly transcends a Darwinian world view.

Experimental evolution can be used in conjunction with biophysical characterization of enzymes to determine the link between cellular fitness and physicochemical properties of enzymes. Sequencing of ancestral and evolved populations can be used to compare the outcomes of experimental evolution with measurements of fitness, using growth rate assays to correlate fitness outcomes to specific mutations. Combined with enzyme assays of kinetic properties that can provide a direct link between genotypic and phenotypic changes of adaptive mutants, we can model the complex relationship between genotypic changes and evolutionary outcomes. Two experimental evolution systems were used to explore the link between enzyme properties and fitness outcomes. In the first series of studies, a “weak link” evolution experiment was used to explore the effect of reducing selection strength on altering accessible pathways for adaptation. In the weak link method the essential gene for adenylate kinase (AK) was replaced in the chromosome of the thermophile Geobacillus stearothermophilus with a homolog from Bacillus subtilis. Replacement with the maladapted gene confers a high fitness cost, and therefore mutations that restore function of AK are strongly favored. Two triple mutants of AK containing a new combination of single point mutants identified under strong selection, AKQ199R/A193V/Q16L and AKQ199R/T179I/Q16L were discovered through an adaptation experiment using a weak temperature ramp; suggesting that the adaptive landscape for AK thermostability is highly constrained. A thermostable coupled assay was developed for measuring adenylate kinase activity using LDHTTHERMOPHILUS and PKGSTEAROTHERMOPHILUS at high temperatures. The triple mutants had increased function compared with the double mutant ancestors, but the triple mutants displayed diminishing returns epistasis on fitness. In the second experimental evolution system, a mathematical model was developed to investigate the role of adaptive mutations, in the tetracycline inactivation enzyme TetX2, on antibiotic resistance to minocycline (MCN). Growth rates measurements, enzyme kinetics, and flux balance equations were used to develop a model to predict the effect on growth rates of TetX2 and seven adaptive TetX2 variants at different MCN concentrations. Population histogram measurements for the experimental evolution study were measured using a high throughput Illumina sequencing method (FREQ-SEQ). We found that the model was able to accurately predict the fitness outcomes for the wild type and the seven single mutants of TetX2 that were originally isolated, as well as for a double mutant that was not used in the development of the original model. The mathematical model accurately predicts that the two mutants TetX2T280A and TetX2N371I provide the largest fitness benefits, in agreement with the results of in vitro experiments on adaptation to MCN. The model was also able to accurately predict enzyme parameters from growth rates values, with a specific emphasis on predicting the ratio of Vmax/KM(MCN). The model allows us to make predictions about the fitness benefits of physicochemical changes to enzymes, and can be used as a high throughput method for determining enzyme kinetic parameters without requiring protein purification. Understanding how physicochemical changes of enzymes relate to phenotypic changes, and ultimately to fitness, requires knowledge of both the molecular basis for determining enzyme properties, and how selection acts on fitness differences to determine evolutionary outcomes. This research provides direct links between physicochemical changes and adaptive phenotypes, as well providing observations of how adaptive landscapes and fitness changes affect evolutionary outcomes.

No description available.

Thesis (Ph.D.)--Virginia Commonwealth University, 2009. / Prepared for: Dept. of Biology. Integrative Life Sciences Program. Title from resource description page Includes bibliographical references.

Thesis (Ph. D.)--University of Queensland, 2001. / Includes bibliographical references.

IRC+10420 is a massive (> 20M☉), very luminous (> 〖10〗^6L☉) star that is in the rare phase of evolution from the red supergiant to the luminous blue variable or Wolf-Rayet phase. Previous observations reveal that the circumstellar envelope is rich in molecular gas, and can be detected out to a radius of about 8” = 6.0 X 〖10〗^17 cm. Observations in CO also reveal that the global massloss rate of IRC+10420 has changed dramatically over the last 6000 years, comprising two major episodes of mass loss lasting for about 1000 and 4000 years respectively separated by period of very low mass-loss rate lasting for about 1000 years. Surprising, previous observation in SiO(J = 2 - 1) revealed a ring-like enhancement at a radius of about 1” (7.5X 〖10〗^16 cm) from the star, contrary to the expectation that SiO molecules should be frozen onto dust grains very close to the star (within ～ 〖10〗^16cm). This ring-like enhancement has been attributed to a large-scale shock produced by interactions between faster and slower moving portions of the expanding envelope. In this thesis, we mapped the circumstellar envelope in SiO(J = 1 - 0) to better constrain the physical conditions (gas density, temperature and SiO abundance) in the SiO-emitting gas. We find a similar ring-like enhancement in SiO(J = 1 - 0) but located further out at a radius of about 2” (1.5 X 〖10〗^17 cm), and confirm that the SiO emission extends as far out as the CO envelope. The computed SiO(J = 2-1)/SiO(J = 1-0) line ratio significantly exceeds unity at radius out to about the location of the ring-like enhancement (2”), and drops to a value of about unity beyond this radius. From a one-dimensional non-local thermodynamic equilibrium model, we explore the physical conditions that can reproduce the observed brightness temperatures in both SiO(J = 10) and SiO(J = 2-1) as well as their line ratio as a function of radius. The SiO-emitting gas is required to have a density that is much higher (from a factor of a few to about two orders of magnitude) than has been inferred for the CO-emitting gas at the same radii. The required surface filling factor of the SiO-emitting gas depends on their unknown gas-phase SiO abundance; for an abundance of ～〖10〗^(-5), as inferred just above the photospheres of lowmass evolved stars, the surface filling factor of these condensations range from ～0.001 to ～0.1. Thus, the SiO emission from the envelope of IRC+10420 most likely originates from dense condensations that are immersed in more diffuse gas that produces the bulk of the observed CO emission. We reason that the SiO-emitting condensations correspond to the dust clumps detected in reflected light with the Hubble Space Telescope. These dust clumps are distributed from near the star out to a radius of 2”, spanning the same extent as the peaks of SiO- (and CO-) emitting envelope. We show that these dust clumps are expanding in every direction away from the stars at a velocity that is significantly higher than the CO-emitting gas, and anticipate that shocks thus generated heats up the dust clumps to release SiO into the gas phase. / published_or_final_version / Physics / Master / Master of Philosophy

published_or_final_version / Physics / Master / Master of Philosophy