Thomas Hahn

Evolutionary adaptations are no longer required to ensure the survival of Homo Sapiens It is assumed that during the RNA world there was no aging but simply replication with errors that allowed for the required evolutionary diversity to adapt to changing environmental conditions. Once the sophistication of replication improved and much more stable proteins started to evolve aging was the only way to adapt the genotypes to changing environmental conditions. Therefore, evolution has developed a highly sophisticated set of adaptations to regulate the aging process and to ensure that it occurs under all possible conditions so that despite the width of genetic variations we have not been able to observe phenotypes, which by chance lack all aspects of aging in most species. But now, we have become so good in adapting and optimizing our environmental conditions to the needs of our genotypes and phenotypes by building houses, planes, roads, computers, antibiotics, etc. that we no longer need evolution because we are in the process to rapidly learning how to create the optimal environmental conditions for our particular genotypes. For example, if I had lived only 10,000 years ago, I'd most likely be killed very early in live since I am visually impaired due to Albinism and hence would not have been able to run away on time from a hungry tiger because I'd see it way too late as in the case of the robbers last night. Or I'd died early of skin cancer since my skin cannot tan causing me to get severe sunburn even only after 10 minutes of sun radiation. But now, at least in certain countries, I can function well despite being affected by two harmful loss of function mutations in my tyrosinase genes, which prevent the synthesis of the brown human pigment melanin, which is required for developing the macular, which is the point of highest vision acuity on the retina and thus causes me to be legally blind and for skin tanning to protect from UV radiation. Today, at least in America, even despite being genetically inferior, I have a chance of earning a PhD in bioinformatics if I can get sufficient access to the needed critical resources. They do already exist but the question is whether I can get access to all of them. Therefore, our task is now to abolish what evolution took so much time to develop within the next 20 years. An indication supporting the feasibility of this approach is that certain knockout mutations can extend lifespan thus putting them in the class of evolutionarily evolved suicide genes that must be switched off. Possible scenarios for the gradual evolution of aging: Let's assume that long time ago during the transition from the RNA into the protein world there first evolved only one lifespan limiting bottleneck in a process essential to maintain life thus increasing the speed of evolution. Then it no longer mattered for how much longer life would be possible considering only the remaining processes essential to maintain life for which no lifespan limiting bottleneck had yet evolved. On the one hand, the emergence of the protein world slowed down the speed of evolution by improving the fidelity of the replication processes but at the same time accelerating it or at least insure a minimum rate of evolution through aging. Therefore, mutations causing bottlenecks in the remaining essential life processes would accumulate since they'd no longer affect the phenotype. How to distinguish between causes and consequences of aging?

Aging could be regulated by the interplay between many different kinds of data-dimensions, all of which provide a fraction of information and dependencies, which must be manipulated in such a way that our evolved internal suicide clock, which is most likely driven by our developmental genes, can not only be stopped but also reversed, because our lives should no longer depend on a kind of evolution, which selects for mechanisms that cause our lifespan to be finite. Long time ago, back in the RNA world, evolution could not select against an individual RNA strand without adversely affecting its replication rate. Because back then, everything, which helped the RNA strand to withstand degradation and stressors, also helped its replication. Hence, there was no distinction between the individual and the replication-relevant material, since both were exactly identical and therefore, they could not be separated. But now evolution can select against individual parents without adversely affecting any relevant aspect of replication. As long as the entire individual was completely composed of exactly the same matter, which was essential for replication, e.g. an individual RNA strand, there was-by default-no aging at all-but instead-only replication. Aging could only evolve in the protein world because then not all the physical matter, of which the parents consisted, was essential for replication anymore. Only this distinction allowed evolution to select for active killing programs, which are most likely driven either directly by actively programmed destruction mechanisms, e.g. apoptosis, or indirectly by neglecting to maintain, repair and restore essential functions, e.g. chaperone-aided protein-folding, peroxisome degradation, or maintaining the steepness of the needed proton-, salinity-, ion-and nutrient-gradients across membranes because our evolved in-built suicide clock killed faster than those life-essential processes declined enough for posing a threat on life. The life-cycle, i.e. the time span from birth to death, seems to be very similar to the cell cycle because it appears to consist of long phases of relative stability and little change interrupted by short periods of rapid changes, which can be as drastic as metamorphosis in species, like worms or frogs, but which nevertheless can be found to a lesser extend in all species. The periodic interval pattern of changes is too similar across members of the same species to be solely the result of the much more randomly acting wear and tear process alone. Women, for example, lose their ability to have children between 50 and 60 years of age. This low variation makes it impossible for this loss of function being caused by wear and tear alone. The same applies to the lifespan. Its variation between members of the same species is way too small for claiming that its length is determined by wear and tear alone. Therefore, I believe that it is likely that there is actually an actively regulated and well timed transition mechanism, which works similar to cell cycle checkpoints, from old age into death.

There have been insurmountable obstacles for sharing instructions with disabled students but now they can be easily overcome by using the free programs TeamViewer 5 in combination with Skype, Krut Computer Recorder and Zoomtext because they allow to transmit all instructional material in accessible format in real time from the lecture computer directly to the screen of the disabled student.

The Program of All-Inclusive Care for the Elderly (PACE)
(2013)[6] study aimed to find out if the program we designed
for the 11 month treatment can efficiently help people lose
weight, and even can keep tracking of weight loss and body fat
by checking some of the parameters we measured during the 11
months. We worked on the potentially significant parameters
for weight loss in 11 months, such as age, height, weight, body
size and body fat. We used association rule mining and
classification rule mining to discover which parameters are
significant for weight loss and what are the associations
between weight loss and those significant parameters.
Experimental results showed that weight loss with support from
0.2 to 0.9 and confidence from 0.7 to 1.0 is related to body
weight and the changes of chest size, arm size, waist size, thigh
size and hip size. In future, we will discover the associations
among body weight, body size, body fat, heart beat and blood
pressure.

The objective is to understand how sterol transport is affected by caloric restriction (CR) and aging.  Sterol synthesis generally increases as cells age.  Yet, despite this increase, the sterol content of selected organelle membranes, such as lysosomal membranes, declines with age and adversely affects cellular functions. We are interested in a better understanding of changes in sterol transport to membranes of vacuoles (the yeast counterparts of mammalian lysosomes) in response to anti-aging and pro-aging manipulations.

The objective is to understand how sterol transport is affected by caloric restriction (CR) and aging.  Sterol synthesis generally increases as cells age.  Yet, despite this increase, the sterol content of selected organelle membranes, such as lysosomal membranes, declines with age and adversely affects cellular functions. We are interested in a better understanding of changes in sterol transport to membranes of vacuoles (the yeast counterparts of mammalian lysosomes) in response to anti-aging and pro-aging manipulations.

Since genes mutate at different rates, a more informative phylogenetic tree can be constructed when faster and slower mutating genes are considered in conjunction. Performing these analyses manually by copy-pasting and aligning one gene sequence after another is very laborious, tedious, time consuming, and prone to significant errors.  To alleviate these problems, we developed an open access web application called “Dione Bioinformatics Framework”. The Dione Bioinformatics Framework is a web-based application that utilizes an intuitive user friendly graphical interface. The first release of the framework includes two main features. 1.) It can concatenate multiple sequences and 2.)  Perform sequence alignments. The Dione Bioinformatics Framework is not command line based but rather allows dragging and dropping the steps of the concatenation and alignment process. These processes are symbolized by icons that can be connected by arrows. The application performs two primary functions, namely "concatenation" and "alignment". Both features can be used together or in isolation to specify which functions are used in any order. Each function is represented as nodes in the workflow, with nodes connected to other nodes by arrows. The direction of the arrow represents the execution order. The application allows users to create multiple workflows and subsequently execute them in a single run.  In future releases, the framework is intended to include all the applications needed for detailed phylogenetic analyses. This application is intended to be an exercise based teaching aid for conveying evolutionary principles to high school and undergraduate students with limited programming background.

Only by stopping and preferably reversing the aging process, immortality is possible.  Aging is the gradually progressing - and so far inevitable - decline of physiological functions eventually leading to death.  Age-related diseases are caused by impairments of essential metabolic pathways to such an extent that they are significantly impairing the vitality of the organism as a whole.  Many age-related diseases are adversely affecting the nervous system and/or cardiovascular system and are often the final causes of death.  Telomerase overexpression has been shown in mouse studies to lead to extended lifespan [Bernades et al]. It has also been shown by Harley et al, that a natural compound TA-65, when taken as a dietary supplement drastically reduced the percentage of senescent cells in individuals. These results come only as recent characterizations of the contents of dried roots of Astralagus used in traditional Chinese medicine. De Jesus et al further showed that the effect of TA-65 is telomerase dependent; in this proposal I am outlining experiments to further characterize the molecular mechanism of TA-65 and determine its ability to help slow down the aging process by enhancing telomerase activity.

Please click here http://singledrug.com/media/Temporal_Analysis_of_Aging_description.pdf to see it. I plan to add the following: I found Osh6 at a very high rank of lifespan extending genes. That is why I want to take all the genes that are part of the same molecular function as Osh6 and run an expression similarity analysis on SPELL (Serial Pattern of Expression Levels Locator) (see spell.yeastgenome.org/). I hope to find genes of unknown functions that could be involved in the same molecular function. If I find this, I'll validate with my time series plots and TFBS similarity analysis. That way I hope to predict the function of an unknown gene in lipid metabolism. I will also try the opposite approach, i.e. I'll take genes of unknown function and look for a group of similarly expressed genes, which belong to the same GO term. I added 19 pages to the second chapter. I was able to further improve the accuracy of my yeast lifespan predicting machine learning algorithm by adding more features about lipids. My last draft had an average prediction error of +/-5 replications. Now it is less than +/-3 replication for the 50th percentile of the replicative lifespan. To my knowledge nobody has published any research on about using machine learning and feature selection to predict the replicative age. If you have any more ideas how I could make my dissertation look better, I'd be very thankful for any new feasible ideas and inspirations. I have mainly been using Python and R for all my analysis. I found that ribosomal RNA and proteins are most closely follow the cell cycle and that if knocked out the yeast's lifespan can be extended significantly. I want to use the different cyclical periodically recurring and highly fluctuating gene expression oscillation pattern to discover the functions of genes we have not yet discovered. I am interested in this because a friend of mine showed me a cancer dataset that looked very similar. Since cancer is very similar to the yeast in that it keeps constantly dividing and hence its age can be expressed in replication we could use the same guilt-by-association approach to infer the functions or at least the way their expression is controlled by TFs from genes with similar temporal expression patterns of genes with already known functions. When analyzing the time series plots of the 62 yeast chaperons I discovered that their expression plot trajectories can easily be clustered into 7 expression patterns, which are consisted with the 7 classes of yeast chaperons. If I had only taken the average centroid of all 62 chaperons I would have been misled because using this approach could not be applied to potentially discover more new currently still unknown chaperons because no unknown chaperon would have time series trajectories, which would highly correlate with the average chaperon expression pattern of all 7 classes since their in-group expression pattern is very homogeneous but between the 7 different chaperon classes the expression pattern is very different. I intended to elucidate the different ways by which we age by a very similar clustering algorithm, which I have developed based on time series curves over three subsequent yeast cell cycle consisting of 36 mRNA measures each about 25 minutes apart; thus, allowing to measure very steep mRNA expression fluctuation periods by manifold within the only 90 minutes of the cell cycle because even very brief but rapid expression changes that last not much longer than 1/12th of a cell cycle, of which I found surprisingly many. I discovered that inferring gene functions based on highly correlated time series curves can only work if mRNA and protein concentration are measured at least 10 times within a cell cycle or else, a very brief but very steep peak or drop in expression could be missed. Any attempt to group

Please click here http://singledrug.com/media/Temporal_Analysis_of_Aging_description.pdf to see it. I plan to add the following: I found Osh6 at a very high rank of lifespan extending genes. That is why I want to take all the genes that are part of the same molecular function as Osh6 and run an expression similarity analysis on SPELL (Serial Pattern of Expression Levels Locator) (see spell.yeastgenome.org/). I hope to find genes of unknown functions that could be involved in the same molecular function. If I find this, I'll validate with my time series plots and TFBS similarity analysis. That way I hope to predict the function of an unknown gene in lipid metabolism. I will also try the opposite approach, i.e. I'll take genes of unknown function and look for a group of similarly expressed genes, which belong to the same GO term. I added 19 pages to the second chapter. I was able to further improve the accuracy of my yeast lifespan predicting machine learning algorithm by adding more features about lipids. My last draft had an average prediction error of +/-5 replications. Now it is less than +/-3 replication for the 50th percentile of the replicative lifespan. To my knowledge nobody has published any research on about using machine learning and feature selection to predict the replicative age. If you have any more ideas how I could make my dissertation look better, I'd be very thankful for any new feasible ideas and inspirations. I have mainly been using Python and R for all my analysis. I found that ribosomal RNA and proteins are most closely follow the cell cycle and that if knocked out the yeast's lifespan can be extended significantly. I want to use the different cyclical periodically recurring and highly fluctuating gene expression oscillation pattern to discover the functions of genes we have not yet discovered. I am interested in this because a friend of mine showed me a cancer dataset that looked very similar. Since cancer is very similar to the yeast in that it keeps constantly dividing and hence its age can be expressed in replication we could use the same guilt-by-association approach to infer the functions or at least the way their expression is controlled by TFs from genes with similar temporal expression patterns of genes with already known functions. When analyzing the time series plots of the 62 yeast chaperons I discovered that their expression plot trajectories can easily be clustered into 7 expression patterns, which are consisted with the 7 classes of yeast chaperons. If I had only taken the average centroid of all 62 chaperons I would have been misled because using this approach could not be applied to potentially discover more new currently still unknown chaperons because no unknown chaperon would have time series trajectories, which would highly correlate with the average chaperon expression pattern of all 7 classes since their in-group expression pattern is very homogeneous but between the 7 different chaperon classes the expression pattern is very different. I intended to elucidate the different ways by which we age by a very similar clustering algorithm, which I have developed based on time series curves over three subsequent yeast cell cycle consisting of 36 mRNA measures each about 25 minutes apart; thus, allowing to measure very steep mRNA expression fluctuation periods by manifold within the only 90 minutes of the cell cycle because even very brief but rapid expression changes that last not much longer than 1/12th of a cell cycle, of which I found surprisingly many. I discovered that inferring gene functions based on highly correlated time series curves can only work if mRNA and protein concentration are measured at least 10 times within a cell cycle or else, a very brief but very steep peak or drop in expression could be missed. Any attempt to group

Aging could be regulated by the interplay between many different kinds of data-dimensions, all of which provide a fraction of information and dependencies, which must be manipulated in such a way that our evolved internal suicide clock, which is most likely driven by our developmental genes, can not only be stopped but also reversed, because our lives should no longer depend on a kind of evolution, which selects for mechanisms that cause our lifespan to be finite.

Long time ago, back in the RNA world, evolution could not select against an individual RNA strand without adversely affecting its replication rate.  Because back then, everything, which helped the RNA strand to withstand degradation and stressors, also helped its replication.  Hence, there was no distinction between the individual and the replication-relevant material, since both were exactly identical and therefore, they could not be separated. 

But now evolution can select against individual parents without adversely affecting any relevant aspect of replication.  As long as the entire individual was completely composed of exactly the same matter, which was essential for replication, e.g. an individual RNA strand, there was - by default - no aging at all - but instead - only replication. 

Aging could only evolve in the protein world because then not all the physical matter, of which the parents consisted, was essential for replication anymore.  Only this distinction allowed evolution to select for active killing programs, which are most likely driven either directly by actively programmed destruction mechanisms, e.g. apoptosis, or indirectly by neglecting to maintain, repair and restore essential functions, e.g. chaperone-aided protein-folding, peroxisome degradation, or maintaining the steepness of the needed proton-, salinity-, ion- and nutrient-gradients across membranes because our evolved in-built suicide clock killed faster than those life-essential processes declined enough for posing a threat on life. 

The life-cycle, i.e. the time span from birth to death, seems to be very similar to the cell cycle because it appears to consist of long phases of relative stability and little change interrupted by short periods of rapid changes, which can be as drastic as metamorphosis in species, like worms or frogs, but which nevertheless can be found to a lesser extend in all species.  The periodic interval pattern of changes is too similar across members of the same species to be solely the result of the much more randomly acting wear and tear process alone. 

Women, for example, lose their ability to have children between 50 and 60 years of age.  This low variation makes it impossible for this loss of function being caused by wear and tear alone.  The same applies to the lifespan.  Its variation between members of the same species is way too small for claiming that its length is determined by wear and tear alone.  Therefore, I believe that it is likely that there is actually an actively regulated and well timed transition mechanism, which works similar to cell cycle checkpoints, from old age into death.

Such kinds of questions are of interest to me and they keep crossing my mind when analyzing time series datasets because they could help to elucidate the mechanisms of aging.  And we must understand them before we can effectively disrupt them. 

We need to start thinking about initiating mechanisms similar to targeted and directed, i.e. intelligently designed and goal-driven evolution, which is aimed at maintaining and restoring all life-essential processes or substituting them accordingly, if they cannot be maintained in the way they have initially evolved.  We need to become fast enough that - if we see a particular approach to fail - we'll still have enough time for quickly developing much better alternatives for preventing the otherwise unavoidable -seeming aging-induced decline, which would inevitably kill us.

############################################################################# Continuously Ongoing Emergency Random Evolution mimicking procedure "Unpredictable Survival" ############################################################################# A major threat is that we are aging much faster than we can reverse it. We are still very far away from inferring, which information is most likely relevant for reversing aging that we MUST take an undirected method to counteract this problem because we don't have any better alternative. Every day lots of new pairs of information is added to the web. Anything, which define at least two indivisible pieces of information as a value pair indicating a specific instance can be evaluated be vsboost. Therefore, we should start developing an independently working software, which keeps crawling the net for any instance defined by at least to informational units as input data. Then, even though this software cannot infer the meaning of any of the event-defining information pair, it can use their values in predicting pretty much any other combination of paired information and try to predict any pair with any other pair. This would allow to identify even weak correlations and dependencies much sooner than when exclusively selecting features manually in our traditional way based on logic reasoning. Although logic reasoning and highly directed and targeted manipulations are good to have it takes us way too much time until our understanding and concepts of new correlations has developed far enough to contribute to logically driven data feature selection and data manipulation. This continuously web-crawling software keeps adding anything, which could either serve as input our output value for any kind of supervised machine learning process. When this software can predict any random feature by whatever means it can possibly think of, it will let us know so we can check whether this could possibly make sense. We need to improve the NLP (Natural Language Processing) and semantic recognizing ability of this randomly feature adding software so that it can combine the same informational components into a singe unit feature. But nevertheless, just like evolution random mistakes in grouping the same information component into a single indivisible feature, variations in the groupings of informational components, which must be predicted at all once, could turn out to be a good thing. For example, considering all TFBS-associated information into a single informational group may allow for the most accurate prediction rate but only when our random model contains all input features we need to define any possible informational dimension needed to sufficiently define all the parameters, which could belong to the TFBS dimension. For example if our feature hungry crawler has not yet discovered that TFBS binding is a cooperative rate than a Boolean process it would fail. But if it could learn to predict time series plots only based on the Boolean value indicating whether a particular TF could possibly bind to a promoter but disregarding the number and order of the TFBS for the same TF in the promoter of one gene it could still predict time series plots well enough to raise its prediction power far above the threshold at which we'd take a look at it. Although this old model is still imperfect it has value to get it asap instead of waiting until our crawler has found all input to parameter to assign a value to all possible dimension of the TFBS domain. This would actually speak in favor of allowing our prediction crawler to randomly vary any specific dimension of any domain suited for training supervised machine learning because the fewer the number of dimensions making up any domain the fewer and smaller information input domain are required for building a model based on randomly considering and randomly grouped information domains. Currently, most of us are not aware of the artificial imperative limitations resulting from letting humans have the monopoly on deciding, which dimensions can be grouped together to form a meaningful instance for input or output to train a supervised model. It is likely that smaller domains consisting of fewer dimensions or larger domain combining more dimensions could be more. But although there are so many humans on this

Why we are still dying despite our understanding of what's needed for reversing aging? By now, 2017, we could have all returned to our teens. For many years already, we all could have been forever young, powerful, healthy, energetic, compassionate, passion, motivated, creative, innovative, flexible and successful. We-i.e. ourselves mainly, could have made the most transforming, meaningful, powerful and happiness-generating difference in our and our loved ones lives if we would no longer be lead by bad and irrational behavior of our most respected role-models because they implicitly convey the fatal misconception that it is still acceptable to keep trying exactly the same methods, which ad utterly failed the last 100 times they were used. This track-record of indisputable failures, loss of lives, griefs and despair still seems to be a widely acceptable justification to fail again by following exactly the same detrimental procedures without considering stopping until 101st failure can no longer be reserved. Below is the beginning of my dissertation. I proposed in May. I must defend in July.. I would like to write such that everyone in this entire world, who knows English, can understand why we-Homo Sapiens-are the only species, which must be blamed and held fully responsible that life-extension, immortality, indefinite youth, happiness, joy still seemed to be far outside of our reach even though they could have played a pivotal, crucial, indefensible role in everyone's life already so long time ago. Death and diseases and suffering could have been some ancient historical concepts of our ancestors. Despite science, medieval concepts of life still seem to predominate. Like 1,000 ago, nobody can escape he inhuman, subconsciously-induced, self-perpetuating, pervasive, omnipresent and never stopping, psychological terror, which inevitably harms, hurts, and haunts everyone indiscriminately. This makes it an invisible prison, which keeps following everywhere.

Research Problem Statement Aging is one of the chief biomedical problems of the 21st century. After decades of basic research on biogerontology (the science of aging), the aging process still remains an enigma. Although hundreds of "theories" on aging have been formulated and many fundamental insights about age-related changes and genetic as well as environmental interventions that change the pace of aging have been discovered, the actual why and how we age remain enigmatic. In the post-genomic era there is an exponential increase in data. As a consequence it is a challenge to utilize all information based on it and derive meaningful knowledge about biological phenomena. No individual scientist, group nor consortium is capable of keeping up within their own field and are overwhelmed by the explosion of data increase. Machine learning applied on biological data has the potential to solve this and cause a paradigm shift from hypothesis-driven research (which predominates biological research including biogerontology) to data-driven research. This dissertation addresses this problem. In particular it proposes and executes the use of machine learning on current existing data to predict drivers of aging (and therefore help to distinguish causes from consequences), interventions to counteract aging, and specific hypotheses to fill in research gaps that require experimental validation. The objective of this project is therefore to build computational models that are based on data relevant to the phenomenon of aging and to predict as many of its aspects and dimensions as possible (thus elucidate their relations to each other). For converting between and sorting within dimensions which are relevant to aging, different machine learning models are evaluated. Ones models are built, it can be determined how much they can explain different aspects of aging. Those models will also be capable of specifying which features are most relevant for prediction (in both classification or regression). It is possible to train models that incorporate age-related changes based on transcriptomic, proteomic, metabolomic, epigenomic as well as morphological data and their combinations. Machine learning is further used to convert between and within them. This work focuses on three types of predictors. Subsequently, discoveries are made with the statistical and learning algorithms. The first model (lifespan predictor) will be trained on predicting the lifespan based on genotype, environment and combinations thereof. It is useful for predicting lifespan extending interventions on the population level. The second model (age predictor) will be trained on predicting the age given features measured on individuals. This one will be useful for identifying biomarkers of aging and to determine the effects of interventions on the level of individuals. The third model will predict functions/regulations of biological entities in regard to the aging process based on heterogeneous data such as ontologies and diverse omics including time-series gene expression profiles (which can be visualized as plots), promoter analysis, and linked data. It will be used to understand the role of genes and proteins as well as perhaps other entities such as small molecules including lipids and other metabolites. In particular it will be attempted to predict the functions of proteins which are still unknown especially those involved in yeast lipid metabolism and its regulation. For this purpose we use primarily yeast as model organism as well as data on humans. Other biomedical model organisms might be added if found beneficial. The predictors can use arbitrary kind of features such as GC content, transcription factor binding energy, transcript and protein length, and various omics among other informative features. A functional interaction network will be build that has static as well as dynamic components. The edges will be derived from experimentally validated physical (protein-protein) and genetic (gene-gene) interaction data, and the size of the nodes will be proportional to significance or protein concentration while the thickness of the edges will be proportional to some metrics (like coexpression correlation) between the nodes connected by that particular edge. Networks will be used to visualize results of predictors. Causal inferences will be based on temporal shifts of correlated profiles.

Please click here http://singledrug.com/media/Temporal_Analysis_of_Aging_description.pdf to see it. I plan to add the following: I found Osh6 at a very high rank of lifespan extending genes. That is why I want to take all the genes that are part of the same molecular function as Osh6 and run an expression similarity analysis on SPELL (Serial Pattern of Expression Levels Locator) (see spell.yeastgenome.org/). I hope to find genes of unknown functions that could be involved in the same molecular function. If I find this, I'll validate with my time series plots and TFBS similarity analysis. That way I hope to predict the function of an unknown gene in lipid metabolism. I will also try the opposite approach, i.e. I'll take genes of unknown function and look for a group of similarly expressed genes, which belong to the same GO term. I added 19 pages to the second chapter. I was able to further improve the accuracy of my yeast lifespan predicting machine learning algorithm by adding more features about lipids. My last draft had an average prediction error of +/-5 replications. Now it is less than +/-3 replication for the 50th percentile of the replicative lifespan. To my knowledge nobody has published any research on about using machine learning and feature selection to predict the replicative age. If you have any more ideas how I could make my dissertation look better, I'd be very thankful for any new feasible ideas and inspirations. I have mainly been using Python and R for all my analysis. I found that ribosomal RNA and proteins are most closely follow the cell cycle and that if knocked out the yeast's lifespan can be extended significantly. I want to use the different cyclical periodically recurring and highly fluctuating gene expression oscillation pattern to discover the functions of genes we have not yet discovered. I am interested in this because a friend of mine showed me a cancer dataset that looked very similar. Since cancer is very similar to the yeast in that it keeps constantly dividing and hence its age can be expressed in replication we could use the same guilt-by-association approach to infer the functions or at least the way their expression is controlled by TFs from genes with similar temporal expression patterns of genes with already known functions. When analyzing the time series plots of the 62 yeast chaperons I discovered that their expression plot trajectories can easily be clustered into 7 expression patterns, which are consisted with the 7 classes of yeast chaperons. If I had only taken the average centroid of all 62 chaperons I would have been misled because using this approach could not be applied to potentially discover more new currently still unknown chaperons because no unknown chaperon would have time series trajectories, which would highly correlate with the average chaperon expression pattern of all 7 classes since their in-group expression pattern is very homogeneous but between the 7 different chaperon classes the expression pattern is very different. I intended to elucidate the different ways by which we age by a very similar clustering algorithm, which I have developed based on time series curves over three subsequent yeast cell cycle consisting of 36 mRNA measures each about 25 minutes apart; thus, allowing to measure very steep mRNA expression fluctuation periods by manifold within the only 90 minutes of the cell cycle because even very brief but rapid expression changes that last not much longer than 1/12th of a cell cycle, of which I found surprisingly many. I discovered that inferring gene functions based on highly correlated time series curves can only work if mRNA and protein concentration are measured at least 10 times within a cell cycle or else, a very brief but very steep peak or drop in expression could be missed. Any attempt to group