Narrow Results

To stimulate interaction and collaboration across scientific fields, we introduce a minimum set of biological criteria that theoretical models of aging should satisfy. We review results of several recent experiments that examined changes in age-specific mortality rates caused by genetic and environmental manipulation. The empirical data from these experiments is then used to test mathematical models of aging from several different disciplines, including molecular biology, reliability theory, physics, and evolutionary biology/population genetics. We find that none of the current models are consistent with all of the published experimental findings. To provide an example of how our criteria might be applied in practice, we develop a new conceptual model of aging that is consistent with our observations.

We consider the Penna model for biological aging to investigate correlations between early fertility and late life survival rates in populations at equilibrium. We consider inherited initial reproduction ages together with a reproduction cost translated in a probability that mother and offspring die at birth, depending on the mother age. For convenient sets of parameters, the equilibrated populations present genetic variability in what regards both genetically programmed death age and initial reproduction age. In the asexual Penna model, a negative correlation between early life fertility and late life survival rates naturally emerges in the stationary solutions. In the sexual Penna model, selection experiments are performed where individuals are sorted by initial reproduction age from the equilibrated populations and the separated populations are evolved independently. After a transient, a negative correlation between early fertility and late age survival rates also emerges in the sense that populations that start reproducing earlier present smaller average genetically programmed death age. These effects appear due to the age structure of populations in the steady state solution of the evolution equations. We claim that the same demographic effects may be playing an important role in selection experiments in the laboratory.

We present a model of evolution of the age structured population based on the Monte Carlo method. We have assumed that the health status of an individual is described by variance of its fluctuations. Each expressed deleterious mutation increases the fluctuations. Additionally, the fluctuations of the environment are superimposed on the fluctuations of individuals in the population. An individual dies if the combination of both stochastic processes trespass the limit (level of homeostasis) set as the model parameter.

The genes are switched on chronologically, what leads to accumulating defective genes expressed during the late periods of life in the genetic pool of the population. That results in the specific age structured population, in accordance with the predictions of Medawar's hypothesis of ageing and the results of the Penna model simulations. A decrease of the variation of the environmental noise increases the average expected lifespan of individuals.

We have adapted the Penna ageing model to simulate the profound changes in the age structures of populations caused by the better life style, medical care and decrease in birth rate. In Poland, after the political transformations in the last decade of the twentieth century, the increase in the expected lifespan has been accompanied by very deep decrease in birthrate, much below the minimum necessary for keeping the constant size of the population. Our microscopic model describes the changes in the age structure which have already happened and predicts the future, assuming that our attitudes in respect to life style and social relations will not change.

This study used reverse selection on populations of Drosophila melanogaster to test the evolutionary theory of aging, including antagonistic pleiotropy and mutation accumulation, the two non-exclusive population genetic mechanisms of aging. Specifically, reversed demographic selection was imposed on five populations selected for late-life fertility for 83 generations (O_1-5), returning them to an ancestral demographic schedule of 14 days. The five ancestral populations (B_1-5) were assayed each generation to serve as a control for environmental fluctuations over time. Relaxing selection for late-fecundity and imposing selection for early fecundity resulted in a rapid drop in longevity, and an increase in early fecundity, suggesting that longevity and some early life fitness component(s) are subject to antagonistic pleiotropy. As longevity fell, the frequency of the S allele of Pgm also decreased. Starvation resistance fell dramatically in reverse-selected males, and remained unchanged in females, suggesting that different physiological genetic mechanisms control resistance to starvation in the two sexes. Desiccation resistance remained unchanged under reverse demographic selection, implicating mutation accumulation as the primary mechanism for its evolution. Overall, these results provide some support for evolutionary theories of aging.

Most deleterious mutations have very slight effects on total fitness, and it has become clear that below a certain fitness effect threshold, such low-impact mutations fail to respond to natural selection. The existence of such a selection threshold suggests that many low-impact deleterious mutations should accumulate continuously, resulting in relentless erosion of genetic information. In this paper, we use numerical simulation to examine this problem of selection threshold.

The objective of this research was to investigate the effect of various biological factors individually and jointly on mutation accumulation in a model human population. For this purpose, we used a recently-developed, biologically-realistic numerical simulation program, Mendel's Accountant. This program introduces new mutations into the population every generation and tracks each mutation through the processes of recombination, gamete formation, mating, and transmission to the new offspring. This method tracks which individuals survive to reproduce after selection, and records the transmission of each surviving mutation every generation. This allows a detailed mechanistic accounting of each mutation that enters and leaves the population over the course of many generations. We term this type of analysis genetic accounting.

Across all reasonable parameters settings, we observed that high impact mutations were selected away with very high efficiency, while very low impact mutations accumulated just as if there was no selection operating. There was always a large transitional zone, wherein mutations with intermediate fitness effects accumulated continuously, but at a lower rate than would occur in the absence of selection. To characterize the accumulation of mutations of different fitness effect we developed a new statistic, selection threshold (ST_d), which is an empirically determined value for a given population. A population's selection threshold is defined as that fitness effect wherein deleterious mutations are accumulating at exactly half the rate expected in the absence of selection. This threshold is mid-way between entirely selectable, and entirely unselectable, mutation effects.

Our investigations reveal that under a very wide range of parameter values, selection thresholds for deleterious mutations are surprisingly high. Our analyses of the selection threshold problem indicate that given even modest levels of noise affecting either the genotype-phenotype relationship or the genotypic fitness-survival-reproduction relationship, accumulation of low-impact mutations continually degrades fitness, and this degradation is far more serious than has been previously acknowledged. Simulations based on recently published values for mutation rate and effect-distribution in humans show a steady decline in fitness that is not even halted by extremely intense selection pressure (12 offspring per female, 10 selectively removed). Indeed, we find that under most realistic circumstances, the large majority of harmful mutations are essentially unaffected by natural selection and continue to accumulate unhindered. This finding has major theoretical implications and raises the question, “What mechanism can preserve the many low-impact nucleotide positions that constitute most of the information within a genome?”

Background. In a companion paper, careful numerical simulation was used to demonstrate that there is a quantifiable selection threshold, below which low-impact deleterious mutations escape purifying selection and, therefore, accumulate without limit. In that study we developed the statistic, ST_d, which is the mid-point of the transition zone between selectable and un-selectable deleterious mutations. We showed that under most natural circumstances, ST_d values are surprisingly high, such that the large majority of all deleterious mutations are un-selectable. Does a similar selection threshold exist for beneficial mutations?

Methods. As in our companion paper we here employ what we describe as genetic accounting to quantify the selection threshold (ST_b) for beneficial mutations, and we study how various biological factors combine to determine its value.

Results. In all experiments that employ biologically reasonable parameters, we observe high ST_b values and a general failure of selection to preferentially amplify the large majority of beneficial mutations. High-impact beneficial mutations strongly interfere with selection for or against all low-impact mutations.

Conclusions. A selection threshold exists for beneficial mutations similar in magnitude to the selection threshold for deleterious ones, but the dynamics of that threshold are different. Our results suggest that for higher eukaryotes, minimal values for ST_b are in the range of 10⁻⁴ to 10⁻³. It appears very likely that most functional nucleotides in a large genome have fractional contributions to fitness much smaller than this. This means that, given our current understanding of how natural selection operates, we cannot explain the origin of the typical functional nucleotide.

There is now abundant evidence that the continuous accumulation of deleterious mutations within natural populations poses a major problem for neo-Darwinian theory. It has been proposed that a viable evolutionary mechanism for halting the accumulation of deleterious mutations might arise if fitness depends primarily on an individual's “mutation-count”. In this paper the hypothetical “ mutation-count mechanism” (MCM) is tested using numerical simulation, to determine the viability of the hypothesis and to determine what biological factors affect the relative efficacy of this mechanism.

The MCM is shown to be very strong when given all the following un-natural conditions: all mutations have an equal effect, low environmental variance, and full truncation selection. Conversely, the MCM effect essentially disappears given any of the following natural conditions: asexual reproduction, or probability selection, or accumulating mutations having a natural distribution of fitness effects covering several orders of magnitude. Realistic levels of environmental variance can also abolish or greatly diminish the MCM effect.

Equal mutation effects when combined with partial truncation (quasi-truncation) can create a moderate MCM effect, but this disappears in the presence of less uniform mutation effects and reasonable levels of environmental variance.

MCM does not appear to occur under most biologically realistic conditions, and so is not a generally applicable evolutionary mechanism. MCM is not generally capable of stopping deleterious mutation accumulation in most natural populations.

The process of deleterious mutation accumulation is influenced by numerous biological factors, including the way in which the accumulating mutations interact with one another. The phenomenon of negative mutation-to-mutation interactions is known as synergistic epistasis (SE). It is widely believed that SE should enhance selective elimination of mutations and thereby diminish the problem of genetic degeneration. We apply numerical simulation to test this commonly expressed assertion.

We find that under biologically realistic conditions, synergistic epistasis exerts little to no discernible influence on mutation accumulation and genetic degeneration. When the synergistic effect is greatly exaggerated, mutation accumulation is not significantly affected, but genetic degeneration accelerates markedly. As the synergistic effect is exaggerated still more, degeneration becomes catastrophic and leads to rapid extinction. Even when conditions are optimized to enhance the SE effect, selection efficiency against deleterious mutation accumulation is not appreciably influenced.

We also evaluated SE using parameters that result in extreme and artificially high selection efficiency (truncation selection and perfect genotypic fitness heritability). Even under these conditions, synergistic epistasis causes accelerated degeneration and only minor reductions in the rate of mutation accumulation.

When we included the effect of linkage within chromosomal segments in our SE analyses, it made degeneration still worse and even interfered with mutation elimination. Our results therefore strongly suggest that commonly held perceptions concerning the role of synergistic epistasis in halting mutation accumulation are not correct.

Loss of information is not always bad. In this paper, we investigate the potential for accelerating the genetic degeneration of RNA viruses as a means for slowing/containing pandemics. It has previously been shown that RNA viruses are vulnerable to lethal mutagenesis (the concept of inducing mutational degeneration in a given pathogen). This has led to the use of lethal mutagenesis as a clinical treatment for eradicating RNA virus from a given infected patient. The present study uses numerical simulation to explore the concept of accelerated mutagenesis as a way to enhance natural genetic attenuation of RNA viral strains at the epidemiological level. This concept is potentially relevant to improved management of pandemics, and may be applicable in certain instances where eradication of certain diseases is sought.

We propose that mutation accumulation is a major factor in the natural attenuation of pathogenic strains of RNA viruses, and that this may contribute to the disappearance of old pathogenic strains and natural cessation of pandemics. We use a numerical simulation program, Mendel's Accountant, to support this model and determine the primary factors that can enhance such degeneration. Our experiments suggest that natural genetic attenuation can be greatly enhanced by implementing three practices. (1) Strategic use of antiviral pharmaceuticals that increase RNA mutagenesis. (2) Improved hygiene to reduce inoculum levels and hence increase genetic bottlenecking. (3) Strategic use of broad-spectrum vaccines that induce partial immunity. In combination, these three practices should profoundly accelerate loss of biological information (attenuation) in RNA viruses.