Molecular Clocks

Does a molecular evolutionary clock exist?  Since the proposal of the molecular clock hypothesis in the 1960¹s, scientists have been pondering that particular question.  This topic still remains one of the most controversial and debated subjects in evolutionary biology.  The molecular clock hypothesis proposes that for any given protein, the rate of molecular evolution is approximately constant over time for all lineages.  Therefore, molecular data can be used for the prediction of time.  Some scientists argue that natural selection produces mutations in genes at such a variable rate that no gene or protein could effectively be used as a molecular clock.  On the other hand, many studies have shown that point mutations can and do occur at relatively regular rates, offering noteworthy evidence for the molecular clock hypothesis.

Problems with the Fossil Record

Oftentimes, paleontologists and molecular biologists particularly disagree on this topic.  The two fields take different approaches to answering evolutionary questions, and therefore, they often come to different conclusions (Hedges 1998).  While the fossil record has provided us with valuable information on phylogeny and divergence times, some say that it is highly biased (Smith 1994).  Some of this bias can be due to the abundance, habitat, or range of a species and sparse sampling (Martin 1993).  Secondly, the fossil record is far from complete.  Large gaps exist in paleontological data; therefore, it is frequently impossible to determine dates of divergence between species.  In addition, oftentimes only a limited number of ³characters² are available for the identification of a species in the fossil record.  This means that some species are named from only a few bones or teeth, and this can lead to erroneous identification.  Lastly, erroneous identification can also result from the fact that species are identified based upon morphology.  This is problematic because species are grouped together by phenotypic traits, or more simply, because they look alike.  However, this does not always indicate that they belong to the same group.

Neutral Theory

Because of these problems, in the 1950¹s, some scientists began looking at molecules in hopes that they could provide us with additional information on phylogeny and divergence times that is absent from or misrepresented in the fossil record.  The Neutral Theory, proposed by Kimura, argues that many amino acid and nucleotide substitutions have little or no functional consequence.  He states that the majority of mutations are inconsequential and therefore, they are not strongly constrained by Natural Selection.  Hence, evolution at the molecular level consists mostly of gradual random replacement of one allele by another that is its functional equivalent.  The neutral mutation rate is therefore constant over evolutionary time because protein function should not alter over time.  Or, in other words, most variation at the molecular level within and among species is effectively neutral and not subject to selection.  Kimura argues that for each protein, evolutionary rate (in terms of amino acid substitutions) is approximately constant per amino acid site per year for various lineages (Nei and Koehn 1983).  Kimura does not, however, claim that all mutations are selectively neutral and he does not claim that selection does not play an important role in shaping the characteristics of organisms.  He proposes only that these cases represent the minority of changes at the molecular level.

The Neutral Theory maintains that for a diploid species with 2N alleles, if the neutral mutation rate per generation is u, then 2Nu represents the number of new mutants arising in the population each generation.  The probability that the allele will become fixed is 1/2N, so the rate of substitution of neutral alleles, k, is equal to the total number of mutations times the probability of their fixation, or

k = (2Nu)(1/2N)

Therefore, k = u (the rate of substitution is equal to the rate of mutation).  Hence, the rate of change should be constant over time and independent of both selection and population size! 


The Molecular Clock Hypothesis

The Molecular Clock Hypothesis is based upon the Neutral Theory.   After the proposal of this theory, scientists began looking at molecules, particularly proteins, to see if Kimura¹s hypothesis would hold true.  From their comparative studies of hemoglobin and cytochrome c protein sequences, Zuckerkandl and Pauling (1962 and 1965) and Margoliash (1963) noticed that rates of amino acid substitution in these proteins were approximately the same among various mammalian lineages (in Li and Graur 1991).  They also noticed that these substitutions were roughly proportional to time as judged against the fossil record.  This led to the proposal of the molecular clock hypothesis by Zuckerkandl and Pauling in 1965.  They hypothesized that for any given protein, the rate of molecular evolution was approximately constant over time in all lineages, or there exists a molecular clock.  According to this hypothesis, molecular data could be used for the prediction of time!  Particularly, molecular data could be used to determine species divergence times and to construct phylogenetic relationships among species, allowing us to fill in the gaps where the fossil record is missing or inadequate.  Their hypothesis proposes that changes in DNA and proteins accumulate at approximately constant rates over geological time.  In this manner, the number of mutations or ³ticks² in DNA and, therefore, the number of substitutions in proteins, is approximately the same per generation for any given organism.  Or, in other words, these molecules change or mutate with clock-like regularity. This theory is also known as the rate-constancy hypothesis.

Evidence for the Molecular Clock Hypothesis

In 1967, Sarich and Wilson presented evidence that serum albumin proteins also change at a regular rate.  From this, they proposed that observed differences in albumins between species could be used as an evolutionary clock to estimate times of divergence and help reconstruct phylogenies (Radinsky 1978).  Futuyama (1986) says that the ³albumin immunological distance² is moderately well correlated with divergence times estimated from paleontological data.  However, there are discrepancies, but this is to be expected when only a single protein is used.

In 1970, cytochrome c was the first protein actually used to estimate divergence time, as this protein was used to create the first phylogenetic tree based on molecular data. Cytochrome c, which is involved in cellular respiration, is an essential enzyme for all species.  Because of this, gene homology has been maintained throughout phylogeny.  The tree was created based on the number of nucleotide base substitutions in the cytochrome c gene for 29 eukaryotic species (Fitch and Margoliash 1970).  Remarkably, the tree matched trees created from paleontological (therefore morphological) data in all but a few points.  This is amazing since it was constructed using only one protein!

Different Proteins Have the Ability to Time Different Evolutionary Events

Based upon molecular clock theory, each particular gene or protein of an organism could possibly serve as a separate molecular clock.  This is based upon the fact that each protein has a distinct rate of evolution depending upon how important its function is.  The less functional constraint on a molecule, the faster it evolves in terms of mutant substitution than those molecules subject to stronger constraint (Nei and Koehn 1983).  For example, histones bind DNA in chromosomes and regulate DNA activity.  Thus, a histone¹s structure is strictly defined because its ability to bind DNA depends upon its particular structure and shape.  The 103 amino acids in this protein are identical for nearly all plants and animals.  It has been one billion years since plants and animals separated, yet of the 103 amino acids in this protein, there is only one difference between a pea and a cow (Zihlman 2001)!  This is due to the enormous amount of functional constraint on this molecule as a result of its essential function.  Conversely, fibrinopeptides can perform their role in blood clotting with almost any amino acid change.  The 20 amino acids in this protein differ by 86% between a horse and a human (Zihlman 2001).  Fibrinopeptides exhibit a very fast rate of change because they¹re subject to less functional constraint.  Because of differing constraints, and, in turn, differing rates of mutation, different molecules can time events in different evolutionary time frames.  For example, histones time once-in-a-billion year events, while fibrinopeptides can clock changes within the past five million years (Zihlman 2001).  Therefore, when doing a study, it is essential to select a molecule that is appropriate to the time span of interest.

Molecular Clocks are Stochastic Clocks

There are two basic types of clocks, metronomic and stochastic.  With the metronomic clock, the distance between ticks is uniform.  In the stochastic clock, the distances fluctuate.  Because of the randomness of nucleotide substitution, it is very unlikely that any molecular clock functions as a metronomic clock.  Until recent years, few, if any, researchers proposed that molecular evolution could be stochastic.  However, it now appears as if some molecular clocks may indeed behave in a stochastic pattern.  Radioactive decay is an example of a stochastic clock.  However, nucleotide substitution has been found to be more random than radioactive decay.  Therefore, amino acid and nucleotide substitutions are not as predictable as radioactive decay for use as ³ticks² in a molecular clock (Fitch 1977).  It has been found that the variance in the time interval between nucleotide substitutions is roughly twice that for a truly stochastic process.  Therefore, to reach the same level of precision, one must count twice as many nucleotide substitutions as radioactive disintegrations.  Fitch states if one averages over a sufficient number of nucleotide substitutions, in different proteins and different species, one can expect reasonable dates of evolutionary divergence.  He says that there really is a clock, but its just unusually erratic (Fitch 1977).

³Erratic Clocks² Yield Disparate Rates

Even supporters of the rate constancy hypothesis acknowledge the fact that molecular clocks can, in fact, behave erratically.  Some genes have been shown to evolve at disparate rates across genes and lineages and over time (Ayala et al.1998).  A great deal of data exists that shows this variation.  One of the most widely used examples are the copper (Cu), zinc (Zn) superoxide dismutase (SOD) and glycerol-3-phosphate dehydrogenase (GPDH) proteins.  These clocks were found to behave in a very erratic manner (Ayala 1997).  Ayala found that with SOD, the rate of amino acid replacement was five times faster among mammals than between fungi and animals.  He also found that these two proteins yielded very different divergence times for animal phyla and the divergence of kingdoms (Ayala 1996).  In addition, the chloroplast genome for six very different plants has been studied and it was found that the rates of evolution are variable from one species to another. However, the investigators say that much of the heterogeneity can be accounted for by differences in generation time (Fitch and Ayala 1994(b)), which will be discussed below.  Also, Goodman and his supporters contend that there are non-uniform rates, despite molecular clock times, in proteins as a result of the acceleration-deceleration pattern of protein evolution, discussed below (Goodman 1981).  Because of time constraints, I have explained only a very few of the clocks that are known to behave erratically.  Many others are known to exist, and some will be discussed below.


Empirical Tests for the Molecular Clock Hypothesis

Futuyama describes the two main empirical tests of the rate-constancy hypothesis that have been used to estimate divergence using molecular information in his book, Evolutionary Biology.  In the first method, the number of nucleotide differences (either determined directly or inferred from amino acid differences) serves as a measure of genetic distance between species.  This number is then plotted against the time of divergence, which is estimated from the fossil record.  A linear correlation suggests a constant rate of divergence per year (Futuyama 1986).  This correlation can then be used as a calibration to estimate the time of divergence where the fossil record is inadequate or absent. 

For the kind of test discussed above (a test often used in molecular studies) a good fossil record is required to calibrate the clock.  But, as described earlier, there are many problems with the fossil record.  Some scientists argue that paleontological data is frequently insufficient to provide accurate estimates of divergence times by which to calibrate the clock (Radinsky 1978 and Kumar and Hedges 1998).  Calibration is, therefore, a very important issue that needs to be addressed.  The dispute over erroneous calibration has resulted in an enormous amount of disagreement between fossils and molecular data and between studies themselves.

Because much of the debate over the molecular clock hypothesis involves conflict over paleontological data, specifically dates of species divergence, Sarich and Wilson (1973) proposed a test that does not require knowledge of divergence times.   Hence, no data from the fossil record is required.  This test is known as the Relative Rate Test.  Futuyama identifies this test as the second major method for testing a constant rate of divergence.  Looking the figure below (from Li and Graur 1991, p. 81), if one wants to compare the rates of molecular substitution in lineages A and B, then a third species, C, will be used as a reference, or an outgroup.  One must be sure that species C branched off earlier than A and B.


The number of substitutions between species A and C, known as KAC, is equivalent to the sum of substitutions that have occurred from point O to point A , KOA, and from O to C, KOC; therefore, KAC = KOC + KOA.  Likewise, KBC = KOB + KOC and KAB = KOA + KOB.  We can now solve the equations to find that:

KOA = (KAC + KAB ­ KBC)/2

KOB = (KAB + KBC ­ KAC)/2

KOC = (KAC + KBC ­ KAB)/2

To determine if the substitution rate is equal in lineages A and B, one would compare the value of KOA with KOB.  If molecules diverge at a constant rate (follow the molecular clock hypothesis), then KOA and KOB should be equivalent, or KOA + KAB = 0.  Also, KOA ­ KOB = KAC ­ KBC, allowing one to compare the rates of substitution in A and B directly from KAC and KBC (Li and Graur 1991). 

While the relative rate test eliminates the need for paleontological data, it only allows one to test the rate-constancy hypothesis for two lineages relative to a third.  It also does not allow one to test for variation within the lineages themselves, or, for example, to detect acceleration or deceleration rates within the lineage.  However, more sophisticated phylogenetic tests allow the clock hypothesis to be tested using many sequences simultaneously.  Three types of phylogenetic tests that are commonly used are the Two-cluster test, the Branch length test and the Log-likelihood ratio test.  Because of time constraints, it is impossible to review all of the phylogenetic tests commonly in use.  Each has its advantages, as well as disadvantages.  No one test is applicable to all situations, and one must be careful to select an appropriate test for the study of interest.

Although the validity of the rate constancy assumption has always been debated, it has been widely used in the estimation of divergence times and in the reconstruction of phylogenetic trees.  Futuyama (1986) says that many of the phylogenetic trees derived from molecular data agree with those created from morphological information.  A good example is the tree that was created based on amino acid sequences of the protein cytochrome c that was previously discussed.  Recall that it resembled the traditional phylogeny in all but a few details (Futuyama 1986).

            Goodman (1981) describes the basic procedure for determining evolutionary rates from the distribution of mutations on the branches of the genealogical trees created from amino acid sequence data when one assumes the rate-constancy or molecular clock hypothesis.  In this method, one first looks at divergence times from the fossil record and chooses an ancestral node to accurately set the clock for the range of dates to be calculated from it.  Secondly, this node¹s time span to the present (million years before present-MyBP) is equated to its number of nucleotide substitutions to the present (its NR span) averaged over the lineages descending from it.  The other nodes are dated by extrapolation from the ratio of each such node¹s NR span over the ancestral node¹s NR span.  Numerous methods exist to aid in molecular phylogenetic reconstruction.  However, again, there is not any one method that works best for all situations.

Statistical Models

Various statistical models are also available.  Therefore, an additional consideration is assigning the proper statistical model to the study.  Some say that more sophisticated statistical models need to be developed for proper calibration of the clock.  Yang (1996) says that simple models do not correct for multiple hits and have caused biased estimation of speciation dates.  Grishin published a set of simple equations that correct for various circumstances, including different rates of change at different sites.  Feng and Doolittle (1997) used these equations to show a linear relationship between calculation distances and the number of allowed mutations based on the observed variation of rate at all sites in various proteins.  However, Grishin¹s equations, and other newly-published models, are not fully understood.  We have yet to find the perfect model to calibrate the clock.

A good example of this problem and the need for more sophisticated models is the previously referenced SOD and GPDH clocks.  Recall that these clocks are frequently used as examples of erratic clock behavior.  These two proteins were originally found to yield disparate divergence times for animal phyla and the divergence of kingdoms.  However, Fitch and Ayala (1994(b)) eventually found that the two proteins ³behave like a fairly accurate clock by assuming a complex pattern in which different sets of amino acids have different probabilities of change that are nevertheless constant through time².  This led them to conclude that ³ (i) the inference that a given gene is a bad clock may sometimes arise through a failure to take all the relevant biology into account and (ii) one should examine the possibility that different subsets of amino acids are evolving at different rates, because otherwise the assumption of a clock may yield erroneous estimates of divergence times on the basis of the observed number of amino acid differences.²  Molecular clocks, therefore, are sometimes much more complex than originally thought, and  new models must be developed and investigated to account for this complexity.

Controversy over the Molecular Clock Hypothesis

The molecular clock hypothesis and Neutrality Theory were met with a great deal of opposition-- with the main contenders being Goodman and his coworkers.  Goodman (1981) and his associates (Czelusniak 1982) argue that proteins evolve at accelerated rates when advantageous mutations are being selected for, particularly at sites that are acquiring functions.  For example, they say that accelerated rates of evolution occurred early in the history of life when proteins were first evolving.  During this time (when advantageous mutations are being selected for), they say that Natural Selection acts as a positive transforming force to improve cell metabolism.  However, he contends that once these new sites are well established, or after extensive types have evolved, that Natural Selection will then act mainly as a stabilizing force.  Or, in other words, Natural Selection then slows these rates (causes deceleration in evolution) by protecting them from further change.  They also contend that there could be additional periods of accelerated rates when new environments are encountered.  Goodman calls this the acceleration-deceleration pattern of vertebrate evolution.  He proposed this theory from his studies of globin evolution.  He and coworkers claim (1974 and 1975) that there was an increased rate of mutant substitution in the early stages of globin evolution, due to the selection of advantageous mutations that improved the function of the protein, followed by a decreased rate during the last 300 million years (Goodman 1981).  Kimura argues, however, that Goodman and coworkers obtained these results only as a result of the erroneous assignment of geologic dates to duplication events (Nei and Koehn 1983).

Goodman (1961) and colleagues (Goodman et al. 1971) have also suggested that a decelerated rate of evolution occurred in hominoids (humans and apes) after they separated from the Old World monkeys.  However, this finding was also met with opposition.  Wilson et al. (1977) argued that the slowdown is an artifact, again stating that Goodman and coworkers assigned an incorrect estimate to a divergence date.  They allege that Goodman assigned a wrong date to the ape-human divergence time.  Wilson et al. performed relative-rate tests using both immunological distance and protein sequence data and concluded that there was no evidence for a hominoid slowdown (Li and Graur 1991).  In an attempt to resolve this controversy, Li et al. (1987) used DNA sequence data and applied the relative rate test to compare the rate of nucleotide substitution in the human and Old World monkey lineages.  They concluded that the Old World monkey lineage evolved approximately 2X faster than the human lineage, offering evidence that the clock does indeed run more slowly in man than apes and monkeys.

Contenders of the Neutralist and Molecular Clock Hypotheses, the Classical Evolutionists, believe in the Selectionist or Darwinian view of evolution.  Darwin proposed that organisms become progressively adapted to their environment by accumulating beneficial mutants (Nei and Koehn 1983).  It seems plausible to advocates of this theory that it should extend to the molecular level.  Supporters of this theory argue that positive selection of advantageous mutations is the driving force behind protein evolution (Goodman 1981). They believe that although mutations do play some role in evolution, Natural Selection is the dominant force that shapes the genetic make-up of populations and that mutations/substitutions occur because of selection for advantageous alleles.

Evolution at the morphological and physiological levels is very erratic.  Hence, classical evolutionists often do not support the molecular clock hypothesis because the suggestion of rate constancy at the molecular level does not mirror what¹s observed morphologically and physiologically.  Selectionists became particularly skeptical when Sarich and Wilson (1967) used the rate-constancy hypothesis to propose that humans and African apes diverged at an estimated 5 million years ago.  This was in stark contrast to the accepted date proposed by paleontologists of at least 15 million years ago (Li and Graur 1991).  Supporters of the molecular clock hypothesis acknowledge the fact that positive Darwinian selection is, in fact, the major cause of evolutionary change at the phenotypic level.  However, they argue that mutation pressure and random genetic drift are the main factors in evolutionary change at the molecular level (Nei and Koehn 1983).

Global Clocks v/s Local Clocks

It is known that some proteins do not make good molecular clocks.  However, a few researchers still believe that certain genes or regions of genes may, in fact, evolve at a constant rate across all lineages and through time, that is there exists a ³global² clock.  While the search for a universal molecular clock continues, many (including myself) believe that one will never be found.  The skeptics say that we will never find a global molecular clock because of organisms¹ differing metabolic rates, DNA repair efficiency, exposure to mutagens, and generation times (Chao and Carr 1993).  For example, it has been found that the same protein can change more rapidly in one species than in another, or that molecular clocks can ³tick² at different rates.  Wu and Li (1985) and Gu and Li (1992) have looked at genes and protein sequences from rodents and man and compared them with mammal outgroups. The rate of synonymous nucleotide substitutions in the rodent lineage was found to be twice that of the human lineage.    Goodman quoted this study as evidence that the rate of evolution of several proteins has declined in the hominoid primates compared to other vertebrates.  It has also been found that the Old World monkey lineage evolved two times faster than the human lineage (Li and Graur 1991).                                                        Britten (1986) hypothesized that the observed differences in substitution rates could be partly attributed to differences in the efficiency of the DNA repair system of the organism.  For example, some data suggests that rodents¹ DNA repair systems may be less effective than those of humans.  Therefore rodents accrue more mutations per replication cycle (Li and Graur 1991).  Secondly, Martin and Palumbi (1993) have shown that organisms with high metabolic rates have higher mutation rates compared to those with lower metabolic rates, possibly as a result of increased rates of DNA synthesis.  Lastly, disparate rates of substitution are also noted between organisms with very different generation times.  Kohne (1970) proposed the generation-time effect to account for this observation.  The higher substitution rates in monkeys than in humans and rodents than in primates could perhaps be explained by this theory (Li and Graur 1991).  The generation-time effect proposes that since the rate-constancy hypothesis is constant with respect to time, if species have different generation times, then their rates of substitution may also vary.  For example, the number of substitutions per million years may be elevated in organisms with shorter generation times compared to those with longer generation times because mutations are thought to occur at an approximately constant rate per generation (Futuyama 1986). 

However, when one compares organisms with similar generation times, such as mice and rats, the rate-constancy hypothesis holds very well (Li and Graur 1991).  So, although the possibility of a universal or global clock has been refuted by many, it seems more probable that ³local² clocks may exist between closely related species because they often have the same generation times, metabolic rates, etc.

Current Topics and The Future of Molecular Clocks

            The applications of molecular clocks are endless.  They can be used for much more than just constructing evolutionary phylogenies and establishing timelines.  Recently, scientists have been looking at molecular clocks in hopes that they will aid in reconstructing the evolution of modern humans, tracing migrations, determining the origins of domestic plants and animals, and identifying the evolution of cultural patterns and behaviors (in both humans and non-humans). 

Many scientists are currently using the concept of molecular clocks to attempt to reconstruct the evolutionary past of modern humans.  Currently, there are two main opposing views about where and when Homo sapiens originated: the ³Out-of-Africa² hypothesis and the ³multiregional² hypothesis.  The ³Out-of-Africa² hypothesis contends that modern humans developed in Africa (from H. erectus) and migrated form there 100,000-200,000 years ago, replacing H. erectus.  According to this theory, all humans today are descended from a single speciation event in Africa (Powledge and Rose 1996).  The ³multiregional² hypothesis proposes that H. erectus populations evolved into modern humans in many regions through interbreeding between populations.  Once again, molecular and paleontological data are in disagreement with one another.  Molecular data tends to support the ³Out-of-Africa² hypothesis, while paleontological data supports the ³multiregional² view.  

A study published in 1987 by Cann, Stoneking, and Wilson supports the ³Out-of-Africa² hypothesis.  Cann et al. used mitochondrial DNA (mtDNA) comparisons of 147 people from Europe, Africa, Asia, Australia, and New Guinea to establish that all present human mtDNA is descended from a single African woman (Powledge and Rose 1996).   mt DNA is inherited only from the mother, therefore no mixing occurs from generation to generation.  Cann et al. were therefore able to assume that all changes in the mtDNA were the result of mutations over time.  From the 147 subjects, they detected 133 mtDNA genotypes.  They subsequently isolated a sequence of mtDNA that all the subjects had in common and assumed that it must have came from a common ancestor to all modern humans.  Then, using the known rate of mutation of mtDNA among other primates, the researchers calculated how much time that ancestral mtDNA would have taken to mutate into the 133 variants.  They traced it back to an African woman whom they claim lived about 200,000 years ago.  They gave her the name ³Eve² (Cann et al. 1987).  A subsequent and more rigorous study seemed to confirm their original hypothesis (Vigilant et al. 1991).

Recently, however, the validity of this study has came into question.  Geneticist Alan Templeton and others have argued that there were statistical and sampling flaws in the study.  Again, the reliability of ³molecular clocks² has been questioned as well as the rate of mutation used in calculating Eve¹s date (Powledge and Rose).  Some scientists are even questioning the assumption that mtDNA is only passed through the female line (Hagelberg 1999 and Eyre-Walker et al 1999).  Additional evidence will be needed in order to reach an acceptable conclusion and to answer age-old questions about our evolutionary history.  Many scientists are currently in the process of using molecular clocks to gain this additional information.  

            Additionally, groups from all over the world are now in the process of studying DNA from the Y chromosome, in hopes of finding ³Adam.²  The Y chromosome is passed only from father to son, and certain portions of it do not recombine with the mother¹s genes.  Therefore, changes in these portions of the  Y chromosome are assumed to have been caused only by mutations, as in mtDNA.  Assuming, again, that all living humans share a common male ancestor (the Out-of-Africa theory), it should be possible to estimate when this so-called ³Adam² lived.  One of the first studies comparing DNA from the Y chromosome was performed by Dorit et al. in 1995.  They concluded that the first modern human male lived some 270,000 years ago, a date relatively consistent with that of ³Eve²(Dorit 1995).  Similar studies have produced comparable results.  Hammer reduced this date to 188,000 years ago, a date highly consistent with the mtDNA evidence for females (Hammer 1995).  And Whitfield et al. (1995) have arrived at an even earlier estimate of 37,000-49,000 ago; but the basic conclusions are all the same‹the ³multi-regional hypothesis² is unsupported based on this data, while the ³Out-of-Africa² hypothesis is supported.

Scientists are also using molecular phylogenies reconstructed from molecular clock data and time scales based on these phylogenies to draw inferences about the evolution of other characters, including behavior patterns of both humans and non-humans.  Peters and Tonkin-Leyhausen (2000) have mapped vocalization types in cats using a published molecular phylogeny of the cat family that was constructed using the clock hypothesis.  They found that the distribution of these behaviors was congruent with the phylogeny.  Therefore, they were able to generate a time frame for the evolution of these vocalizations.  Interestingly, they observed a period of stasis for several million years in particular vocalization types. Based upon this, they questioned the prevailing hypothesis that behavioral characters are more susceptible to evolutionary change than morphological ones.  Likewise, some scientists are also looking at clocks to reconstruct the spread of language.  For example, Chinese geneticist Du Ruofu has been collecting samples of mtDNA from modern populations to attempt to determine the spread of Indo-European language throughout China.

The data derived from molecular clocks is also being used to reconstruct the evolution of  human cultural practices.  Bryan Sykes has used DNA analysis to question the principal theory about the spread of agriculture into Europe.  Most believe that it was people practicing agriculture who spread into Europe, rather than the idea of agriculture.  However, using mtDNA analysis, Sykes claims that the ancestors of most modern Europeans arrived at least 20,000 years ago, long before the supposed arrival of Neolithic farmers.  He analyzed mtDNA from more than 800 modern Europeans and concluded that the indigenous hunter-gatherer population remained intact and learned how to farm (Powledge and Rose 1996). 

            Nonhuman DNA also has great potential for shedding light on cultural practices.  Luikart, et al. (1998) used mtDNA sequences from 80 breeds of goats to conclude that three main lineages of goat breeds appear to have diverged 200,000-300,000 years ago, long before animals were first domesticated.  This led them to conclude that goats were domesticated from local wild animals at least three times.  They say that a similar pattern exists in cattle, sheep, and pigs.  They found that, in goats, only 10% of the mtDNA variation was partitioned among continents.  However, this value was 50% for cattle.  They conclude that this ³suggests extensive intercontinental transportation of goats and has intriguing implications about the importance of goats in historical human migrations and commerce.² 

            Studies based upon molecular clocks are constantly producing new data, which supports or refutes old hypotheses, as well as leads to new hypotheses.  New data will, no doubt, shed light upon our past, as well as the pasts of non-humans. 


Though the general concept of molecular clocks is widely accepted, controversy still exists over the particulars, especially accurate calibration.  Calibration is an essential piece missing from the puzzle and will be difficult to establish, given the controversy that exists between paleontologists and molecular biologists over the fossil record and the inadequacies in this paleontological data.  Many scientists hope that in the future we will see paleontological data and molecular data as complementary forces to open the doors to the past, rather than the current competing forces.  And while they disagree about most things, both proponents and opponents of the molecular clock theory seem to agree that individual proteins do not make good clocks.  However, when the rates of several different proteins are averaged for remote events, most agree that this method can provide acceptable data to determine dates of divergence.  Even Goodman has admitted that ³when results from different proteins are combined, the hypothesized protein clock does not perform too badlyŠ²  And Fitch stated that ³thus the clock is significantly erratic, but averaged over enough events, the variability tends to be washed out.² 

Lastly, the search for a universal molecular clock as well as local clocks continues.  New gene sequences are becoming available by the day.  The availability of gene sequences will allow us to determine if a global clock and/or local clocks really exist and to perhaps solve one of the most heavily debated questions in evolutionary biology: Does a molecular clock really exist?  I believe that reliable molecular clocks do exist and once the particulars are worked out, the applications of molecular clocks are endless.  I believe that they can and will be used to unlock the mysteries of the past.


Ayala, F.J.  1996.  Molecular clock or erratic evolution? A tale of two genes.  PNAS 93, Oct. 1996:11729-


Ayala, F.J.  1997.  Vagaries of the molecular clock.  PNAS 97, July 1997:7776-83.

Ayala, F.J., A. Rzhetsky, and F.J. Ayala.  1998.  Origin of the metazoan phyla:  Molecular clocks

confirm paleontological studies.  Proc. Natl. Acad. Sci. USA.  95:606-611.

Britten, R.J.  1986.  Rates of DNA sequence evolution differ between taxonomic groups.  Science


Cann, R.L., M. Stoneking, and A.C. Wilson.  1987.  Mitochondrial DNA and human evolution.  Nature


Chao, L., and D.E. Carr.  1993.  The molecular clock and the relationship between population size and

generation time.  Evolution 47:688-90.

Czelusniak, J.M., D. Goodman, M.L. Hewett-Emmett, P.J. Weiss, Venta and R.E. Tahian.  1982. 

Phylogenetic origins and adaptive evolution of avian and mammalian haemoglobin genes.  Nature 298:297-300.

Dorit, R.L., H. Akashi and W. Gilbert.  1995.  Absence of Polymorphism at the ZFY locus on the human Y chromosome.  Science 268:1183-1185.

Eyre-Walker, A.., N.H. Smith, and J. Maynard Smith.  How clonal are human mitochondria?  Proc. Royal

Society B, 1999, Vol.266, pp.477-483.

Feng, D. and R. Doolittle.  1997.  Converting amino acid alignment scores into measures of evolutionary

time.  J. of Molecular Evolution 44(997):361-70.

Fitch, W.M.  1977.  The phylogenetic interpretation of macromolecular sequence information: simple

methods.  In M.K. Hecht, P.C. Goody and B.M. Hecht (eds.) Major patterns in vertebrate evolution, pp. 169-204.  Plenum Press, New York.

Fitch, W.M. and F.J. Ayala.  1994(a).  The Superoxide Dismutase Molecular Clock Revisited.  PNAS


Fitch, W.M. and F.J. Ayala.  1994(b).  Tempo and mode in evolution.  PNAS 91, July 1994:6717-6720.

Fitch, W.M. and E. Margoliash.  1970.  The usefulness of amino acid and nucleotide sequences in

evolutionary studies.  Evol. Biol. 4:67-109.

Futuyama, D.J.  1986.  Evolutionary Biology. Sinauer Associates, Sunderland, Mass.

Goodman, M.  1981.  Decoding the pattern of protein evolution.  Prog. Biophys. Mol. Biol. 38:105-64.

Goodman, M.  1961.  The role of immunochemical differences in the phyletic development of human

behavior.  Hum. Biol.  33:131-162.

Goodman, M., J. Barnabus, G. Mausuda, and G.W. Moor.  1971.  Molecular evolution in the descent of

man.  Nature.  233:604-613.

Goodman, M., G.W. Moore, J. Barnabas, and G. Matsuda.  1974.  J. Mol. Evol. 3:1.

Goodman, M., G.W. Moore, and G. Matsuda.  1975.  Nature.  253:603.

Gu, X. and W.H. Li.  1992.  Higher rates of amino acid substitution in rodents than in humans.  Mol.

            Phylogenet. Evol. 1(3):211-14.

Hagelberg, E., N. Goldman, P. Liu, S. Whelan, W. Schiefenholven, J.B. Clegg, D.K. Bowden. 1999. 

Evidence for mitochondrial DNA recombination in a human population of island Melanesia. Proc

R Soc Lond 266: 485-492.

Hammer, M.F.  1995.  A recent common ancestry for human Y chromosomes. Nature 378:376-378.

Hedges, B.  1998.  Molecules and the Fossil Record.  From Nature Debates.  Macmillan Publishers Ltd,


Kohne, D.E.  1970.  Evolution of higher-organism DNA.  Quart. Rev. Biophys.  33:327-375.

Kumar, S. and S.B. Hedges.  1998.  A molecular timescale for vertebrate evolution.  Nature 392:917-920.

Li, W.H. and D. Graur.  1991.  Fundamentals of Molecular Evolution.  Sinauer Associates, Inc.,

Sunderland, Mass.

Li, W.H., M. Tanimura, and P.M. Sharp.  1987.  An evaluation of the molecular clock hypothesis using

mammalian DNA sequences.  J. Mol. Evol.  25:330-342.

Margoliash, E.  1963.  Primary structure and evolution of cytochrome c.  Proc. Natl. Acad. Sci. USA 


Martin, R.D.  1993.  Primate origins: plugging the gaps.  Nature 363: 223-234.

Martin, A.P. and S.R. Palumbi.  1993.  Body size, metabolic rate, generation time, and the molecular

            clock.  Proc. Natl. Acad. Sci. USA 90(9):4087-91.

Nei, M. and R.K. Koehn (eds.)  1983.  Evolution of Genes and Proteins.  Sinauer Associates, Sunderland,


Peters, G. and B. Tonkin-Leyhausen. 2000.  The tempo and mode of evolution of acoustic communication

signals of felids.  Evolution of Communication 2: 233 - 248.

Powledge, T. M., and M. Rose.  1996.  The great DNA hunt.  Archaeology 49(5).

Radinsky, L.  1978.  Do albumin clocks run on time?  Science 200:1182-3.

Sarich, V.M, and A.C. Wilson.  1973.  Generation time and genomic evolution in primates.  Science


Sarich, V.M, and A.C. Wilson.  1967.  Immunological time scale for hominid evolution.  Science 


Smith, A.B.  Systematics and the fossil record.  Blackwell Scientific Publications, London, 1994.

Vigilant L., M. Stoneking, H. Harpending, K. Hawkes, and A.C. Wilson.  1991. African populations and

the evolution of human mitochondrial DNA. Science 253:1503-1507.

Whitfield, L.S., J.E. Sulston, P.N. Goodfellow.  1995.  Sequence variation of the human Y chromosome.

Nature 378:379-380.

Wilson, A.C., Carlson, S.S., and White, T.J.  1977.  Ann. Rev. Biochem. 46:573.

Wu, C. I., and W. H. Li.  1985.  Evidence for higher rates of nucleotide substitution in rodents than in man. 

Proc. Natl. Acad. Sci. USA 82:1741-45.

Yang, Z.  1996.  Among-site rate variation and its impact on phylogenetic analysis.  TREE 11:367-372.

Zihlman, A.L.  The Human Evolution Coloring Book.  Coloring Concepts Inc., New York: Harper Collins,


Zuckerkandl, E., and L. Pauling.  1965.  Evolutionary divergence and convergence in proteins. 

pp. 97-166.  In V. Bryson and H. J. Vogel (eds.), Evolving genes and proteins.  Academic Press, New York.

Zuckerkandl, E, and L. Pauling.  1962.  Molecular disease, evolution and genic heterogeneity. 

pp. 189-225.  In M. Kasha and B. Pullman (eds.), Horizons in Biochemistry.  Academic Press, New York.