Sci. Aging Knowl. Environ., 16 March 2005
Vol. 2005, Issue 11, p. pe7
[DOI: 10.1126/sageke.2005.11.pe7]


The Genetic Basis of Aging: An Evolutionary Biologist's Perspective

David N. Reznick

David N. Reznick is at the Department of Biology, University of California, Riverside, CA 92521, USA. E-mail: david.reznick{at}

Key Words: evolutionary theory of aging • convergence • parallelism • polygenic trait • laboratory culture • natural population • genetic basis of aging • aging-related genes


From the perspective of an evolutionary biologist, one of the most impressive and surprising results to emerge from recent work on the genetics of aging is the finding that homologous genes and biochemical pathways play an important role in determining life span in a wide diversity of organisms. George Williams (1), who was one of the first individuals to develop a formal theory of why senescence evolves (see Williams Classic Paper), explicitly addressed the goal of seeking genes that influence aging. He cautioned that we would be tempted to do so but that we should not waste our time because there are none to be found. He said this because he, and Medawar (2) before him, thought that senescence was either a negative byproduct of the evolution of traits that enhance fitness early in life or was a consequence of the random accumulation of deleterious gene mutations that act late in life (see "Aging Research Grows Up"). They predicted that genes that emerge as important for senescence in any given lineage would be random, so there is no reason to expect any consistency among species in the genes that contribute to senescence. This, at least, was the expectation. What we are finding instead is that there are homologous genes, such as those associated with the insulin/insulin-like growth factor-1 signaling pathway, that are now known to contribute substantially to life span in organisms as diverse as Caenorhabditis elegans, Drosophila melanogaster, and Mus musculus (3-5). Given the taxonomic scope of these model systems, it seems reasonable to argue that such genes may be general to all Animalia in the role that they play in determining life span. How can we reconcile such results with Williams' expectations? Although we cannot offer a full reconciliation in a short essay, we can at least consider what these results might mean from the perspective of an evolutionary biologist.

The first part of our argument is built around the realization that life span is a complex polygenic trait. If we compare the development of such traits in different organisms, then knowing, for example, that some regulatory genes play a common role in very different organisms yields limited information about the underlying genetic architecture, because there is a diversity of downstream genes that may be regulated and because there is potential overlap with other regulatory genes. Showing that individual regulatory genes play similar roles in distantly related organisms thus yields knowledge about only a small fragment of the genetic basis of the trait. We will expand on this argument with recent examples that illustrate the diversity of genetic pathways that can yield similar traits.

The second part of our argument addresses the nature of the organisms that have been used to study the genetics of aging. Well-established model organisms that have been bred in the laboratory are used for such genetic studies. We will argue that these laboratory model organisms have been subjected to similar conditions in laboratory culture that can lead to inadvertent artificial selection. It is possible that some of the similarities among these organisms are a byproduct of such selection.

Concepts from Evolutionary Biology

We begin with a brief primer on some concepts from evolutionary biology that are integral to these arguments. One concept is called "convergence" and is well exemplified in Fig. 1. This figure illustrates two species of lizards that look alike and share very similar lifestyles. Both are well camouflaged in their natural background, are relatively slow moving, and tend to use their camouflage to hide from predators, rather than run away. Both specialize in eating ants. They also have similar patterns of reproduction. When we see such phenomena, we ask "why are these animals so similar?" There are two general answers. One possibility is that they inherited these similarities from a common ancestor. The other is that they did not inherit them from a common ancestor but evolved them independently of one another as they adapted to similar environments. The latter phenomenon is what we refer to as convergence. In this case, we know we are looking at an example of convergence because one lizard, a thorny devil (Moloch horridus), is from western Australia and the other, a horned toad (Phrynosoma cornutum), is from the southwestern United States. These lizards are from different families. Close relatives to each of them have very different appearances and lifestyles. Their similarities are a consequence of each of them independently adapting to similar environments in corners of the world that are very distant from one another.

View larger version (103K):
[in this window]
[in a new window]
Fig. 1. The thorny devil (M. horridus) from western Australia (top) and the Texas horned lizard (P. cornutum) from the southwestern United States (bottom). The similarities in morphology, diet, and patterns of reproduction in these lizards evolved independently in each as they adapted to similar environments. Their similarity is an example of evolutionary convergence. [Photo of M. horridus, credit T. Garland/University of California, Riverside; photo of P. cornutum, credit Texas Parks and Wildlife]

"Parallelism" is a concept related to convergence. It again refers to populations of organisms that evolve similar adaptations independently of one another, but parallelism differs from convergence in the nature of the starting point. Convergence is usually reserved for different and often distantly related species that adapt to similar environments. Parallelism is reserved for different populations of the same species that are isolated from one another but adapt in similar ways to similar environments. The distinction between convergence and parallelism is often made because it is assumed in parallelism that the underlying genetic architecture will be the same, whereas in convergence it may not be the same.

Convergence and parallelism are important concepts for our argument because they can apply to regulatory pathways as well as to morphology. For example, the apparent predominance of insulin or insulin-like signaling pathways in aging in nematodes, flies, and mice are implicitly interpreted as having been inherited from a common ancestor; however, we submit that they might represent convergent evolution.

The Genetic Architecture of Complex Traits

We consider two recent examples from the literature that show that the same trait can be arrived at by different genetic pathways, even when starting with different populations of the same species or with closely related species.

In the first example, Lenski and coworkers (6) used Escherichia coli as a model organism to study the genetics of adaptation. These researchers began with 12 replicate populations (Fig. 2), all of which were derived from a single population that was in turn derived from a single E. coli cell.

View larger version (14K):
[in this window]
[in a new window]
Fig. 2. Simplified diagram of Lenski's selection experiment on E. coli. Twelve replicate lines, all derived from a colony initiated from a single cell, underwent selection for 20,000 generations on a new medium. The evolution of fitness (measured as increased population growth rate relative to that of the common ancestor) was evaluated (6).

Each of the 12 replicates was subjected to the same type of selection because they were all cultured in the same glucose-limited minimal medium. This medium was different from the medium that their common ancestor and its ancestors had adapted to for many thousands of generations before the experiment. The 12 replicates were cultured on the new medium for 20,000 generations. An attractive feature of bacteria is that it is possible to store the ancestor and the evolving lineages in a freezer so that they can be thawed and then compared with one another at a later date. The trait that was evaluated was the capacity of the 12 selection lines to increase in population size relative to their common ancestor when the two were combined and reared together on the new medium. All 12 replicates evolved a more rapid rate of increase in population size relative to their common ancestor, presumably because they were better able to use their new resource base. This evolution of competitive ability is an example of parallel evolution because the 12 replicates were derived from a common ancestor and were then exposed to the same type of selection, but evolved independently of one another because they remained separate for the 20,000-generation duration of the experiment.

Cooper et al. (6) used microarrays to characterize differences in gene expression between two of the lineages and the common ancestor. The parallel evolution of competitiveness in the two lineages was matched by significant parallel changes in the expression of 59 genes. In one of the two lineages, the authors identified a single mutation at the spoT locus that was responsible for the changed level of expression in a subset of the 59 genes. (SpoT functions to control the intracellular concentration of guanosine tetraphosphate, an effector molecule that regulates the expression of many genes.) An allelic replacement of the wild-type spoT gene in the ancestor with the mutant form caused a 9% increase in replication rate and the same pattern of change in the expression of most members of the subset of genes, so there is good evidence for a cause-and-effect relationship between this mutation and adaptation to the new medium. However, the spoT mutation was absent in the second lineage, in spite of both lineages having the same pattern of gene expression changes in the 59 genes. Furthermore, the pattern of gene expression in the second lineage was unchanged if its spoT allele was replaced with the mutant form from the first lineage. Their conclusion was that the two lineages evolved parallel changes in replication rate and gene expression, but did so via different regulatory pathways. Thus, in this example, the combination of downstream genes that were directly responsible for the trait was the same in the two lineages, whereas the upstream part of the pathway differed in the two lineages.

We consider this study important because it is the simplest example of evolution that we can imagine. Cooper et al. initiated the experiment with replicates that were as close to being genetically identical as possible, then exposed them to identical selection. They showed that the phenotypes evolved in the same way in each of the replicates. If ever we could expect a deterministic, uniform genetic architecture to underlie the evolution of a polygenic trait, then this is it, yet there is clear diversity in the way that regulatory pathways evolved as each replicate adapted to the new culture medium.

Gompel and Carroll (7) present a second example that shows that different genetic pathways can lead to the same trait. These investigators used the diversity of pigment patterns seen among species of drosophilid flies as a model system for studying the evolution of pigmentation. Their starting point was to combine the various pigment patterns seen on the bodies of different species in this genus (Fig. 3) with a phylogeny, or family tree, that describes the patterns of relatedness among species of Drosophila. Their tree is a composite of information derived from the morphology of the flies and the similarity in DNA base pair sequences for several loci. A first step was to map the pigmentation patterns of each species onto the family tree to identify species that are likely to be similar in color pattern because they inherited this pattern from a common ancestor versus those that are similar as a consequence of convergent evolution, meaning that they independently evolved similar pigment patterns. For example, D. buskii and D. willistoni are closely related and are similar in having transverse bands of pigment and lacking black tips of the abdomen, probably because they inherited this pattern from a common ancestor. D. melanogaster and D. pallida also have similar pigment patterns, save for the absence of sexual dimorphism in D. pallida. Furthermore, they both have closer relatives that lack such pigmentation. This sort of pattern suggests that they may have evolved similar pigment patterns independently of one another.

View larger version (39K):
[in this window]
[in a new window]
Fig. 3. The phylogeny for species in the genus Drosophila studied by Gompel and Carroll. On the left is the "family tree," constructed with morphological and DNA sequence data. Each branch of the tree represents a different species. Pictured is the appearance of the abdomen, with males and females presented separately if there is a sexual dimorphism in pigmentation. The middle shaded diagram outlines the distribution of trichomes on the abdomen, whereas the diagrams on the right illustrate the pattern of bab gene expression (shown in green). [Reproduced from (7) with permission from Nature]

To investigate the underlying genetic architecture of pigment pattern formation, Gompel and Carroll evaluated the role of candidate regulatory genes in governing the development of these pigment patterns in different species. Earlier research on D. melanogaster established that the distribution of pigment is regulated by the bric a brac (bab) gene. This gene actually regulates both the development of pigmentation and trichome formation. Trichomes are small, hairlike sensory structures found on the bodies of drosophilids. The expression of bab appears to repress the development of pigmentation and promote the development of trichomes, so there is an inverse distribution of trichomes and pigmentation on the abdomen of D. melanogaster. bab plays a similar dual role in many of the species that independently evolved similar patterns of pigmentation and trichome distribution. However, there are also exceptions. For example, D. serrata, which is closely related to D. melanogaster, has the same pattern of expression of bab as D. melanogaster and the same pattern of trichome development, yet the sexual dimorphism in pigmentation is reversed; in D. melanogaster, males have a black tip on the abdomen, whereas in D. serrata, females display this phenotype. D. santomea is also closely related to D. melanogaster and also shares the same pattern of bab expression and trichome distribution in males, but differs from D. melanogaster in its pigmentation. Gompel and Carroll conclude that bab regulates the development of pigment and trichomes in most species but has been disassociated from the development of pigment in these two close relatives of D. melanogaster. Thus, a gene that regulates the expression of a trait in one species is not necessarily responsible for the same trait even in a closely related species. Again, there can be a disassociation between regulatory genes and the traits that they regulate, or (said differently) the same trait can be arrived at via different genetic pathways in different organisms.

A different perspective on the potential complexity of the genetic architecture of life span is presented by Murphy et al. (8). They used two methods to identify genes in C. elegans whose level of expression was significantly up- or down-regulated by DAF-16, a forkhead transcription factor that functions in the insulin/insulin-like growth factor-1 pathway and influences life span (see Johnson Subfield History, Sonntag and Ramsey Perspective, and Larsen Perspective). Murphy et al. then used RNA interference (RNAi) to characterize the function of each downstream gene. Down-regulation of over 300 of these genes by RNAi had a significant impact on life span. This list included genes that play a role in stress resistance, function in resistance to bacterial infections, or affect the rate of turnover of metabolites and specific proteins. The expression of each downstream gene is also likely to be affected by multiple transcription regulators, and each transcription regulator affects the expression of potentially hundreds of genes, so a complex web lies between each candidate regulatory gene and the ultimate phenotype.

How do these examples relate to the observation that life span in organisms as distantly related as C. elegans, D. melanogaster, and M. musculus is strongly influenced by the same candidate genes? Although it is possible that these similarities suggest a uniform genetic architecture of life span and a trait that is inherited from a common ancestor, an alternative is that these similarities only represent components of the complex genetic architecture that underlies the development and evolution of a complex trait. It certainly suggests that some components of the genetic architecture may have been inherited from a common ancestor, but there is also considerable room for convergent evolution and diversity in the underlying genetic mechanisms. Wray and Abouhief (9) present an analogous argument for the relation between regulatory genes and the traits whose development is regulated. They argue that, in spite of the growing literature that demonstrates associations between regulatory genes and morphological structures in distantly related organisms, it can also be shown that regulatory genes and traits can be readily disassociated from one another.

Our general conclusion from these examples is that different genetic loci or networks of loci can produce the same phenotype. The loci that play an important role can vary among replicates within an experiment, populations within a species, and among species. Our general inference from this sort of diversity is that any lab study that evaluates the genetic basis of a complex trait in a model organism and that characterizes the role of candidate genes yields "an" answer to how that trait develops, rather than "the one and only" answer.

Our next example represents a shift from the evaluation of candidate genes to a scan for any locus that has a significant impact on life span. In addition, it is closer to home because it deals with the genetic basis of variation in life span in D. melanogaster (10). Leips and Mackay used quantitative trait locus (QTL) analysis to scan the Drosophila genome for loci that are associated with life span (see Service Perspective and Mackay Perspective). Pictured here (Fig. 4) are the QTLs identified on chromosome 3 in three different experiments. The experiments differed in whether they were done on males or females and in the nature of the larval or adult environment. One result is that they readily identified many loci that were associated with life span. Another is that whether or not an individual locus had a significant effect on life span depended on sex and rearing environment. This approach leads to an unbiased survey of genes that are involved in a trait of interest such as increased life span, but with two limitations. First, it only defines regions of chromosomes that are linked to one or more genes that have segregating alleles that influence life span. Second, it will only identify loci that have a sufficiently large impact on a trait to exceed some statistical threshold. Thus, this method will only identify a subset of the loci that contribute to the development of the trait in question. An important conclusion from this unbiased scanning of different D. melanogaster samples is that the loci that play an important role in a trait can vary among individuals as a function of their sex or their environment.

View larger version (9K):
[in this window]
[in a new window]
Fig. 4. QTLs on chromosome 3 in Drosophila that have a significant impact on longevity. The map positions of significant QTLs for life span are shown for 18 treatment groups from three experiments. There were from one to five significant QTLs in each experiment, and many experiments differed in the QTLs that emerged as significant. The numbers on the x axis correspond to the following treatments: 1, virgin males; 2, virgin females; 3, virgin males; 4, virgin males after 30 min of heat shock; 5, virgin males, 29°C; 6, virgin males, 14°C; 7, virgin males, starvation; 8, virgin females; 9, virgin females after 30 min of heat shock; 10, virgin females, 31°C; 11, virgin females, 14°C; 12, virgin females, starvation; 13, virgin males from low larval density; 14, virgin males from high larval density; 15, virgin females from low larval density; 16, virgin females from high larval density. See (10) for further details. [Reproduced from (10) with permission]

Experiments that explicitly incorporate biological diversity or a diversity of environments are required to reveal a diversity of mechanisms. Therefore, any lab study that evaluates the genetic basis of a complex trait in a model organism and characterizes the role of candidate genes is yielding an overly simplified view of the genetic basis of the trait.

The Model Organism Bias

Another possible explanation for why homologous candidate genes and biochemical pathways can play a similar role in determining life span in a diversity of organisms lies in the nature of the study organisms. The model organisms used in these life-span studies are represented by long-term lab-adapted strains that have been inadvertently "domesticated," or subjected to adaptation to the laboratory environment. One measure of their being modified by lab culture is to compare their performance with conspecifics that were recently derived from natural populations. Miller et al. (11) found that newly established colonies of mice derived from natural populations in Idaho had life spans that equaled or exceeded the longest-lived lab lines (see "Give Me Liberty or Give Me an Early Death"). Likewise, Linnen et al. (12) found that newly established populations of D. melanogaster had life spans that were comparable to lab lines that had been selected for extended life span (see Spencer and Promislow Perspective). For both Drosophila and mice, this means that beginning with a lab stock and selecting for increased life span only restores the life spans seen in nature, or that the established lab stocks were shorter-lived than their wild-caught progenitors. Finally, Walker et al. (13) and Jenkins et al. (14) found that newly established lab lines of C. elegans were competitively superior to the longest-lived lab lineages when the two were compared in an environment that mimicked nature (see "Get Wild").

Why are lab lines short-lived or competitively inferior? Domestication, or adaptation to lab culture, generally includes inadvertent selection for short generation times and/or high fecundity in the presence of abundant resources and a relatively uniform environment. For example, Drosophila are often cultured with a 2-week cycle. Sgro and Partridge (15) showed experimentally that such a culture regime selects for an increase in fecundity early in life, but decreased fecundity late in life and decreased life span. The lab culture of any organism should tend to select for such evolution, because individuals with such traits will contribute proportionately more offspring to the next generation. Because the diversity of organisms that have been used to evaluate the role of candidate genes in longevity share this type of lab culture, the similar role played by candidate genes may reflect convergence, or a common way of adapting to lab culture. Specifically, the insulin/insulin-like signaling pathway has been linked to many general functions associated with early development and reproduction, as reviewed by Murphy et al. (8), so its prominence in association with life span in lab-adapted organisms could conceivably be a byproduct of intense selection on early growth, developmental rate, and reproduction. Alternatively, Promislow and Tatar (16) have argued that such laboratory culture will also encourage the accumulation of deleterious mutations that act late in life as a byproduct of the short life spans that are enforced by laboratory culture.

Readers might have a hard time seeing lab culture as being selective or evolution as being fast or pervasive enough to make that much of a difference in an organism's biology. A recent example from the domestication of salmon offers a concrete demonstration of the process (17) (Fig. 5). There has been a decline in egg size of over 30% in domestic salmon over an 11-year period. Age at maturity ranges from 2 to 7 years in this species and varies among stocks; the number of generations represented in this figure likely ranges from two to five. This change is an inadvertent byproduct of a breeding program that was intended to be nonselective and to retain genetic diversity. Eggs and sperm are stripped from adults and then the eggs are fertilized en masse with sperm from several males. The young are reared in a food-rich environment to insure high survival. A consequence of this form of culture is that it removes the usual tradeoff that exists between the number and size of eggs that a female produces in nature. In a natural setting, where food can be limiting, selection favors the production of fewer, larger eggs that produce hatchlings that have extra fat reserves (18). In artificial culture, where food is abundant, there is not the same advantage associated with being large, so there is inadvertent selection in favor of females that produce a greater number of smaller eggs. Such females contribute more offspring to the next generation. Similar selection occurs in all model organisms. The model organisms used to study the genetics of life span and aging have been in culture for many more generations than these salmon, so the potential adaptation to the lab can be that much more extensive.

View larger version (58K):
[in this window]
[in a new window]
Fig. 5. The evolution of egg size under domestication in Chinook salmon. [Graph reproduced in slightly modified form from (17); photo credit D. Heath/University of Windsor]

So what can be done to eliminate biases that may be leading to the discovery that the same genes are important for life span in several model organisms? First, one can expand on the theme of working on organisms that have been recently derived from nature. Second, one can work with natural populations that have been shown to have naturally evolved differences in life span and ask whether the candidate genes or metabolic pathways that have proven to have a large impact on life span in lab lines play a similar role when differences in longevity evolve in a natural setting. The virtue of a natural setting is that such differences in longevity will have evolved under different constraints than in the lab, such as the tradeoff between egg size and egg number in the salmon example.

Whereas most aging-related research has been based on a limited number of lab strains of model organisms, we do know of a few species for which there are natural populations that are known to differ in life span and patterns of senescence. These populations offer the opportunity to evaluate the generality of genetic mechanisms identified in model organisms. We think that the greatest potential lies in species of microcrustacea in the genus Daphnia. There are now good descriptions of dramatic differences in longevity and senescence among populations (19). These differences appear to have evolved repeatedly in response to differences among habitats in the stability of the environment. For example, populations from temporary woodland ponds have a median life span of 37 days as compared with a median life span of 62 days for populations from permanent lakes. Those from temporary ponds have higher juvenile growth rates and higher fitness early in life, but fecundity declines more rapidly with age in comparison to that of lake populations. Daphnia are small and relatively cheap to culture, and standardized methods of culture have been developed. There is a rich literature on the life histories of these organisms and there are well-developed methods for studying the ecology and demography of these organisms in nature [(19) and references therein]. Finally, genomics resources are being developed for Daphnia (20). A second possibility is guppies (Poecilia reticulata), which are now known to also have substantial among-population variation in the rate of senescence as well as in many other traits (21) (see "Prey for Long Life"). Although these fish are not as cheaply cultured as Daphnia and have longer life spans (a maximum life span of over 1400 days), they do offer the promise of a new vertebrate system for the study of senescence. In addition, they can be cultured in the same kind of arrangement used for zebrafish.


Although we are impressed with the finding that homologous genes can extend life span in organisms as different as nematodes, flies, and mice, we are not yet convinced that these similarities represent traits that were inherited from a common ancestor and that they are indices of similarities in the genetic architecture of the aging process. The similar results attributed to homologous genes may be the product of convergent evolution in a restricted number of model organisms that have adapted to similar laboratory conditions. The extent to which these results generalize to organisms adapted to a natural environment have not been evaluated. Even if these similarities really are common to all animals, they have thus far been evaluated in a biased fashion. The initial discovery of candidate genes is often the product of an unbiased screening process, but the investigation of a candidate in new organisms is not. The relative importance of a given candidate gene or pathway in determining a trait relative to all of the other factors that may play a role cannot be ascertained, because only the candidates are considered. The number of downstream genes controlled by a regulatory pathway is potentially large, and the relation between regulatory genes and those that are regulated is flexible. When we reconsider Williams' (1) caution in light of these new results, it is clear that he was partly wrong, in the sense that there really have been genes found that have an affect on the aging process that are common to a diversity of organisms. However, Williams' conclusion that senescence is a consequence of the random accumulation of deleterious gene action late in life, and hence that the causes of senescence will vary from organism to organism, has yet to be fairly tested. Evaluating Williams' proposition requires an expansion of the range of organisms that are evaluated, the range of environments in which they are evaluated, and the range of genetic mechanisms that are considered.

March 16, 2005
  1. G. C. Williams, Pleiotropy, natural selection, and the evolution of senescence. Evolution 11, 398-411 (1957).[CrossRef]
  2. P. B. Medawar, An Unsolved Problem of Biology (H. K. Lewis, London, 1952), p. 24.
  3. M. Holzenberger, J. Dupont, B. Ducos, P. Leneuve, A. Geloen, P. C. Even, P. Cervera, Y. Le Bouc, IGF-1 receptor regulates lifespan and resistance to oxidative stress in mice. Nature 421, 182-187 (2003).[CrossRef][Medline]
  4. M. Tatar, A. Kopelman, D. Epstein, M. P. Tu , C. M. Yin, R. S. Garofalo, A mutant Drosophila insulin receptor homolog that extends life-span and impairs neuroendocrine function. Science 292, 107-110 (2001).[Abstract/Free Full Text]
  5. C. Kenyon, J. Chang, E. Gensch, A. Rudner, R. Tabtiang, A C. elegans mutant that lives twice as long as wild type. Nature 366, 461-464 (1993).[CrossRef][Medline]
  6. T. F. Cooper, D. E. Rozen, R. E. Lenski, Parallel changes in gene expression after 20,000 generations of evolution in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 100, 1072-1077 (2003).[Abstract/Free Full Text]
  7. N. Gompel, S. B. Carroll, Genetic mechanisms and constraints governing the evolution of correlated traits in drosophilid flies. Nature 424, 931-935 (2003).[CrossRef][Medline]
  8. C. T. Murphy, S. A. McCarroll, C. I. Bargmann, A. Fraser, R. S. Kamath, J. Ahringer, H. Li, C. Kenyon, Genes that act downstream of DAF-16 to influence the lifespan of Caenorhabditis elegans. Nature 424, 277-283 (2003).[CrossRef][Medline]
  9. G. A. Wray, E. Abouheif, When is homology not homology? Curr. Opin. Genet. Dev. 8, 675-680 (1998).[CrossRef][Medline]
  10. J. Leips, T. F. C. Mackay, The complex genetic architecture of Drosophila life span. Exp. Aging Res. 28, 361-390 (2002).[CrossRef][Medline]
  11. R. A. Miller, J. M. Harper, R. C. Dysko, S. J. Durkee, S. N. Austad, Longer life spans and delayed maturation in wild-derived mice. Exp. Biol. Med. (Maywood) 227, 500-508 (2002).[Abstract/Free Full Text]
  12. C. Linnen, M. Tatar, D. E. L. Promislow, Cultural artifacts: A comparison of senescence in natural, lab-adapted and artificially selected lines of Drosophila melanogaster. Evol. Ecol. Res. 3, 877-888 (2001).
  13. D. W. Walker, G. McColl, N. L. Jenkins, J. Harris, G. J. Lithgow, Evolution of lifespan in C. elegans. Nature 405, 296-297 (2000).[CrossRef][Medline]
  14. N. L. Jenkins, G. McColl, G. J. Lithgow, Fitness cost of extended lifespan in Caenorhabditis elegans. Proc. R. Soc. London Ser. B 271, 2523-2526 (2004).[Medline]
  15. C. M. Sgro, L. Partridge, Evolutionary responses of the life history of wild-caught Drosophila melanogaster to two standard methods of laboratory culture. Am. Nat. 156, 341-353 (2000).[CrossRef]
  16. D. E. Promislow, M. Tatar, Mutation and senescence: Where genetics and demography meet. Genetica 102-103, 299-314 (1998).[CrossRef][Medline]
  17. D. D. Heath, J. W. Heath, C. A. Bryden, R. M. Johnson, C. W. Fox, Rapid evolution of egg size in captive salmon. Science 299, 1738-1740 (2003).[Abstract/Free Full Text]
  18. S. A. Berkeley, C. Chapman, S. M. Sogard, Maternal age as a determinant of larval growth and survival in a marine fish, Sebastes melanops. Ecology 85, 1258-1264 (2004).[CrossRef]
  19. J. Dudycha, A. Tessier, Natural genetic variation of life span, reproduction, and juvenile growth in Daphnia. Evolution 53, 1744-1756 (1999).[CrossRef]
  20. J. Dudycha, personal communication.
  21. D. N. Reznick, M. J. Bryant, D. Roff, C. K. Ghalambor, D. E. Ghalambor, Effect of extrinsic mortality on the evolution of senescence in guppies. Nature 431, 1095-1099 (2004).[CrossRef][Medline]
Citation: D. N. Reznick, The Genetic Basis of Aging: An Evolutionary Biologist's Perspective. Sci. Aging Knowl. Environ. 2005 (11), pe7 (2005).

Science of Aging Knowledge Environment. ISSN 1539-6150