6
HUMAN EVOLUTION Vol. 2 - N. 5 (407-412) - 1987 P. Darlu Groupe de recherches de g~n~tique @id~miologique, INSERM, U 155, ChAteau de Longchamp, Bois de Boulogne, 75016 Paris P. Tassy INSERM, U 155 and "Laboratoire de Palgontologie des vert~brds et de Pal~ontologie humaine (U.A. 720 du C.N.R.S.), Universitd P. & M. Curie, 4 place Jussieu 75252, Paris Cedex 05 Key words: Phylogeny, mtDNA, Modern Humans Roots (a Comment on the Evolution of human mitochondrial DNA and the Origins of Modern Humans) Recent studies on the evolution of human nuclear DNA and mitochondrial DNA lead to striking conclusionson the Africans origins of modernhumans. Yet, uncertaintiescan be foundin the phylogenetic interpretation for the data. In his comments about CANN, STONEKING• WILSON'S (1987: 31) article on the evolution of human mitochondrial DNA (mtDNA), WAINSCOAT(1987: 13) wrote that <~Eve was alive, well and probably living in Africa around 200,000 years agog>. This statement summarizes the results of recent analyses on nuclear DNA by WAINSCOAT et al. (1986: 491) and on mtDNA by Cann and her colleagues (op. cit.). The first cited paper aroused many comments anct responses (JONES & ROUHANI, 1986a: 449; GILES & AM- BROSE, 1986: 21; WAINSCOAT, HILL & CLEGG 1986: 22; JONES & ROUHANI, 1986b: 599; HILLIS, 1986: 208). Surely, CANN, STONEKING& WILSON'S results will have the same impact. Phylogenetic statements based on human biochemical evolution must be welcomed. Yet, phylogenetic inferences on the evolution of one species based only on intraspecific data represent one of the most difficult goals of phylogenetic research. Previous results obtained by WAINSCOAT et al. (1986) after analyses on nuclear DNA were criticized on methodological grounds. Apparently the phylogenetic analysis by restriction mapping of mitochondrial DNA by CANN, STONEKING & WILSON (op. cit.) has not the flaws that had been found in the article by WAINSCOAT et al. (1986), already noticed by GILES & AMBROSE (1986) and HILLIS (1986) and admitted by WAINSCOAT, HILL & GLEGG(1986), It is enough to recall that WAINSCOAT et al. (1986) used a phenetic clustering method (single linkage) to obtain a dendrogram representing the genetic distance between eight populations based on their frequencies of various 8 globin gene cluster haplotypes. The dendrogram shows <~a large genetic distance.., between the African and non-African populations~> so that it was reinforcing the idea that <cafounder population migrated from Africa and subsequently gave rise to all non-African popula- tions~> (WAINSCOAT et al., 1986: 493). As underlined by GILES ~ AMBROSE(1986), this kind of dendrogram cannot be interpreted in phylogenetic terms. For that matter, without a specification of a root, a dendrogram gives only a representation of similarity distances and does not yield phylogenetic information. It could be said with no risk of conceptual error that phylogenetic inferences can only

Roots (a comment on the evolution of human mitochondrial DNA and the origins of modern humans)

  • Upload
    p-darlu

  • View
    214

  • Download
    0

Embed Size (px)

Citation preview

HUMAN EVOLUTION Vol. 2 - N. 5 (407-412) - 1987

P. Darlu Groupe de recherches de g~n~tique @id~miologique, INSERM, U 155, ChAteau de Longchamp, Bois de Boulogne, 75016 Paris

P. Tassy INSERM, U 155 and "Laboratoire de Palgontologie des vert~brds et de Pal~ontologie humaine (U.A. 720 du C.N.R.S.), Universitd P. & M. Curie, 4 place Jussieu 75252, Paris Cedex 05

Key words: Phylogeny, mtDNA, Modern Humans

Roots (a Comment on the Evolution of human mitochondrial DNA and the Origins of Modern Humans)

Recent studies on the evolution of human nuclear DNA and mitochondrial DNA lead to striking conclusions on the Africans origins of modern humans. Yet, uncertainties can be found in the phylogenetic interpretation for the data.

In his comments about CANN, STONEKING • WILSON'S (1987: 31) article on the evolution of human mitochondrial DNA (mtDNA), WAINSCOAT (1987: 13) wrote that <~Eve was alive, well and probably living in Africa around 200,000 years agog>. This statement summarizes the results of recent analyses on nuclear DNA by WAINSCOAT et al. (1986: 491) and on mtDNA by Cann and her colleagues (op. cit.). The first cited paper aroused many comments anct responses (JONES & ROUHANI, 1986a: 449; GILES & AM- BROSE, 1986: 21; WAINSCOAT, HILL & CLEGG 1986: 22; JONES & ROUHANI, 1986b: 599; HILLIS, 1986: 208). Surely, CANN, STONEKING & WILSON'S results will have the same impact.

Phylogenetic statements based on human biochemical evolution must be welcomed. Yet, phylogenetic inferences on the evolution of one species based only on intraspecific data represent one of the most difficult goals of phylogenetic research.

Previous results obtained by WAINSCOAT et al. (1986) after analyses on nuclear DNA were criticized on methodological grounds. Apparently the phylogenetic analysis by restriction mapping of mitochondrial DNA by CANN, STONEKING & WILSON (op. cit.) has not the flaws that had been found in the article by WAINSCOAT e t al. (1986), already noticed by GILES & AMBROSE (1986) and HILLIS (1986) and admitted by WAINSCOAT, HILL & GLEGG (1986), It is enough to recall that WAINSCOAT et al. (1986) used a phenetic clustering method (single linkage) to obtain a dendrogram representing the genetic distance between eight populations based on their frequencies of various 8 globin gene cluster haplotypes. The dendrogram shows <~a large genetic distance.., between the African and non-African populations~> so that it was reinforcing the idea that <ca founder population migrated from Africa and subsequently gave rise to all non-African popula- tions~> (WAINSCOAT et al., 1986: 493). As underlined by GILES ~ AMBROSE (1986), this kind of dendrogram cannot be interpreted in phylogenetic terms. For that matter, without a specification of a root, a dendrogram gives only a representation of similarity distances and does not yield phylogenetic information.

It could be said with no risk of conceptual error that phylogenetic inferences can only

408 DARLU and TASSY

be drawn from those trees which can be labelled <<phylogenetic trees>>. The term <<tree>> has long been used for various kinds of schemes showing topographic relations between objects, OTU' s or taxa. The vague definition of such a term and its different uses are the cause of great confusion in the evolutionary literature. The cladists have shown that to be phylogenetic, a tree must be rooted and based on a data matrix which gives information on the polarity of data, that is the transformation states of characters from ancestral (= primitive = plesiomorphic) to derived (= apomorphic). Those two conditions are intrinsi- cally tied since the root corresponds to the ancestor which, by definition, shows the ancestral condition for all characters in the data matrix. As has been often said, HENNICS (1966) great achievement was to clearly demonstrate that global similarity could be misleading in phylogenetic terms and that genealogical relationships between taxa can only be infered by shared derived characters (synapomorphies).

The most usual criterion for giving a polarity to character states is the out-group criterion not only in biological studies without ontogenetic information but also in palaeontology. The out-group criterion unanimously used by cladists can be simply defined as follows: if a character state seen in the group under study is also seen outside the group it is primitive for the group under study. In this way, the polarity of characters can be deduced by outgroup comparisons (see, among others, WILEY, 1981: 110, 139; FARraS, 1982: 329).

A large data matrix can only be analyzed by computers. Computed phylogenies based on the parsimony criterion are generally rooted by an out-group. Even if we do not know per se in a DNA sequence of bases if one base is primitive and the other one is derived, the out-group criterion gives the polarity. The phylogeny of the human mtDNA given by CANN, STONEKmC & WILSON (1987, fig. 3: 34) is explicitely called as such: <<genealogical tree for 134 types of human mtDNA>>. Indeed the dendrogram is rooted by an ancestral point. The examination of the tree shows a first dichotomy between Africans on the one hand and all other populations plus some Africans on the other. Apparently, as long as the logic behind placement of the root is not critically examined, this dichotomy is evidence of an early cleavage in the history of modern humans and, as WAmSCOAT writes (1987: 13) this <<evolutionary tree itself., clearly suggests an African origim> (for Homo sapiens sapiens).

The place of origin of modern humans must have been somewhere in the Old World, and why not in Africa? We do not argue against the African hypothesis which is also shared by some palaeoanthropologists (RIGHTMIRE, I979, 1984; BRAtJER, 1984) but not all (VANDEI:MEERSCH, 1982, WOI.I'OFF, WU & THORNE, 1984). Acknowledging the considera- ble improvement brought by Cann and her colleagues to human biochemical evolution, we would just comment here, from a methodological viewpoint, that their <<genealogical tree>> could lead to ambiguous interpretations and brings up different questions:

i) We consider the rooting technique used by Cann, Stoneking and Wilson not to be devoid of ambiguity for their tree cannot be considered as a phylogenetic, or genealogical, tree.

CANN, STONEKING & WILSON (1987: 34) have published a parsimonious tree using the computer program PAUP designed by D. Swofford. The rooting technique was done by the <<midpoint>> option; out-group or ancestor option (using ancestral states of the char- acters) being not applicable to their data. The midpoint option is a mathematical artifice for rooting: the origin of the tree is placed at the midpoint of the longest path connecting two lineages inside the group under study. However trees obtained from the same data but relying on different rooting procedures, as midpoint and out-group, can sometimes provide conflicting results.

COMMENT ON MTDNA AND ORIGINS OF MODERN HUMANS 409

TABLE 1 - Data matrix ]or the cladogram in Figure 1. Apomorphic states are coded 1, plesiomorphic states are coded O, and unknown conditions are coded

hypanc A B C D E F

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 1 1 1 1 1 7 7 ? 0 1 7 7 7 7 0 0 0 7 ? ? ? ? ? ? ? ? ? 7 0 0 0 7 7 7 1 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 ? 0 0 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 1 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 ? 1 ? 1 1 1 ? 7 0 ? 0 0 0 1 1 1 1 1 0 0 0 0 0 0 1 0 1 0 0 1 0 0 0 ? ? ? 0 1 0 0 0 0 ? 1 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 0

In Figure 1 we give the different trees produced by PAUP with three rooting options: rooting by ancestral states, by out-group and by midpoint. All trees are derived from the same data matrix (Table 1). In this example, the ancestor and the out-group bear the same character states (all plesiomorphic). The two trees produced by ~midpoint~> (Figure 1 b-c) are different from the unique tree produced either by ancestor or out-group (Figure la). Even if the ancestor is introduced in the analysis, the midpoint option does not produce a tree similar to the other trees (Figure ld). In this case the midpoint option fails to identify the ancestor as such.

These conflicting results are not really surprising since the midpoint option is not only based on grouping by synapomorphies but also on the amount of differences occuring along the branches. It can be noticed that in our Figure I (based on Table I) the phylogenies show heterogeneous rates of divergence expressed by the different lengths of the sister -branches (this example is not theoretical: it is taken from a cladistic analysis of interrelationships of a mammalian group). It is not surprising then to find discrepancies between the trees rooted by <~midpoint~> (Figure 1 b-c-d) and the phylogenetic tree of Figure la. The midpoint procedure can be an estimate for rooting only if <dt is presumed that the two most divergent OTUs of S (= a collection of OTUs) had the same average rate of evolutionary change~> (FARraS, 1972: 657). Here again we are left with the debate on homogeneous or heterogeneous rates of evolution and on the molecular clock, since this kind of clustering method could be somewhat comparable to phylogenetic method only if we assume constant evolution. Critiques of this model had been raised on different grounds, methodology (FARRIS, 1981) and data (BRITTEN, 1986; LI & TANIMURA, 1987).

Without an out-group (which has to be a taxon outside the human species, such as an African ape, as HILLIS emphasized (1986) after WAINSCOAT et al.'s (1986) paper, it is not possible to be sure that the dendrogram given by Cann, Stoneking & Wilson is a phylogenetic tree. Hence the phylogenetic conclusions raised from it may be questionable.

ii) the numerous bush-like connections along the tree, imply to us that their tree (CANN, STONEFaNC & WILSON, 1987, figure 3) is probably a consensus tree, derived by summing up all trees with equal parsimony, but retaining uncertainties on the relationship between mtDNA types and on the sequential events leading from one type to another.

iii) The authors use a percentage of sequence divergence to scale the lengths of the tree, drawing equal length between each derived mtDNA type and their common ances- tor. Could we conclude that the authors assumed a constant evolutionary rate, not only between the most distant types as performed by the midpoint option, but also over all branches, unless they actually printed out a <~cladistio> tree showing only the branching pattern and not a tree displaying also the length of the branches. If so, we cannot figure out the meaning of the sequence divergence scale.

iv) The authors ignore every site present in only one type of mtDNA. So they leave

410 DARLU and TASSY

a

hypanc

[-Lc F

~ B

I o

b

[ - ' - -~ E

[3

C

C I 1 I E

g 'C

d

[ B

[3

h y p a n c

A

Figure 1 - - Interrelationships of six taxa (A-F) based on the data matrix displayed by Table 1. Parsimony analyses by PAUP (program designed by D. Swofford) with different rooting options.

One unique phylogenetic tree depicting relationships between taxa A, B, C, D, E, F, (la) is rooted either by the out-group option or by the ancestor option (hypanc = hypothetical ancestor). Here all characters of the out-group or the ancestor are plesiomorphic. On Figure 1 b-c, two different trees equally parsimonious are produced with rooting by the <<midpoinb> option. The two trees display an arrangement different from that of tree seen on Figure la. On Figure l-d, the tree is rooted by the <<midpoint>> option with one taxon added: the ancestor. With this rooting option, the ancestor does not appear in an ancestral position but is situated inside one subgroup. In this example, the ~<midpoinb> rooting option gives inadequate results.

COMMENT ON MTDNA AND ORIGINS OF MODERN HUMANS 411

AFR~ICAN []

CAUCA..qlA~ []

AGIAN

AUSTRALIAN

NEW GUISEA~ I

k Figure 2 -- Dendrogram of five modern human populations obtained by average linkage method and based on data provided by CANN et al. (1987, Table 1, upper part).

out 102 among 195 polymorphic sites. By doing such a selection on the sites available for analysis, they reduce to some extent the estimation of the sequence divergence % by a quantity depending on how many and where such sites occur on the tree.

v) The authors represent in a table the mtDNA divergence within and between 5 human populations. Unfortunately, these data cannot be compared to others data drawn from mtDNA types only. For example., the divergence estimated within individuals from New Guinea (CANN, STONEKING & WILSON, 19871 Table 1) is certainly lower than the same divergence estimated within mtDNA different types only, because several New Guineans, sampled in a small area, share identical mtDNA types. This is not true in their African population which has been sampled among Black Americans probably derived from different biological populations settled in a larger area than New Guinea. This comment is just to underline the difficulty of inferring relationships between populations from data which has not been gathered according to a rigourous sampling method. The only conclusion which can be safely drawn from such data concerns the phylogeny of mtDNA types as expressed by Cann, Stoneking & Wilson's in their figure 3, not the phylogeny of the populations bearing such types.

Besides we can get a very different history of the location of ~Edem> when using the % sequence divergence matrix provided by Cann, Stoneking & Wilson and applying an UPGMA or an ~average linkago> clustering method (an unsuitable method in this case). It yields Figure 2 implying that New guinean and African are equally far from their common ancestor, not only in space as we know but also in time so that we could imagine the <~Garden of Eden>> located somewhere between Africa and New Guinea.

Many other results could also be as well provided by the data as long as any relevant hypothesis on evolutionary rate along the branches cannot be put forward with enough confidence and as long as selective pressure, admixture, migration, bottleneck effects cannot be included in intraspecific phylogenetic models. Note also that any such conclu- sions are based on screening about 9% of the mtDNA which itself is only 0,048% of the whole human genome.

In conclusion, we make a short digression by discussing the recent attempts to locate

412 DARLU and TASSY

the centers of origin of Man and by evoking NELSON & PLATNICK'S comments (1981; 357- 375) which have shown convincingly (at least for us) that the search for center of origin was historically, and somewhat conceptually, linked to the search of Paradise. Indeed, the jacket of their book Systematics and Biogeography reproduces a map published in 1868 by Ernst Haeckel depicting the migrations of the <<12 races of Mare> from Paradise situated in Lemuria (in place of the Indian ocean). Nelson and Platnick also considered with their usual provocative style that in modern biogeographic studies the search for centers of origin still was more a metaphysic quest than a testable scientific activity. It is interesting to view the metaphors of the African Eve and the Garden of Eden drawn from CANN, STONEKING • WILSON'S (1987) study, as a <r of those metaphysical concepts, even if they are mainly used for their mediatic strength 1.

References

BRAUER G., 1984. A craniological approach to the origin of anatomically modem Homo sapiens in Africa and implications for the appearance of modem Europeans. In: (F. H. Smith & F. Spencer, eds.). The origins of modern humans: a world survey of the fossil evidence, pp. 327-410. Alan R. Liss.

BRITTEN R. J., 1986. Rates of DNA Sequence evolution differ between taxonomic groups. Science, 231: 1393- 1398.

CANN R. L., STONEKING M. & WILSON A. C., 1987. Mitochondrial DNA and human evolution. Nature, 325: 31-36.

FARRIS J. S., 1972. Estimating phylogenetic trees from distance matrices. American Naturalist, 106: 645-668. FARRIS J. S., 1981. Distance data in phylogenetic analysis. In: (V. A. Funk & D. R. Brooks eds.). Advances

in Cladistics, pp. 3-23. New York Botanical Garden. FAI~RIS J. S., 1982. Outgroups and parsimony. Systematic Zoology, 31: 328-334. GILES E. & AMBROSE S.H., 1986. Are we all out of Africa? Nature, 322: 21-22. HENNm W., 1966. Phylogenetic Systematics. University of Illinois Press, Urbana. HILLIS D.M., 1986. Out of Africa - - through a genetic bottleneck. Nature, 323: 208. JONES J. S. & ROUHANI S., 1986a. How small was the bottleneck. Nature, 319: 449-450. JONES J. S. & ROUHANI S., 1986b, Mankind's genetic bottleneck. Nature, 322: 599-600. LI W. -H. & TANIMURA M., 1987. The molecular clock runs more slowly in man than in apes and monkeys.

Nature, vol. 326: 93-96. NELSON G. & PLATNICK N., 1981. Systematics and Biogeography. Columbia University Press, New York. RIGHTMIRE G.P., 1979. Implications of Border Cave skeletal remains for the later pleistocene human

evolution. Current Anthropology, 20: 25-35. RmHTMIRE G. P., 1984. Homo sapiens in Sub-Saharan Africa. In: (F. H. Smith & Spencer, eds.). The

origins of modern humans: a world survey of the fossil evidence, pp. 295-325. Alan R, Liss. VANDERMEERSCH B., 1982. The first Homo sapiens sapiens in the Near East. In: (A. Ronen ed.). The

transition from Lower to Middle Palaeolithic and the origin of modern man, pp. 297-301. Bar International Series 151.

WAINSCOAT J. S., HILL A. V. S., BOYCE A. L., FLINT J., HERNANDEZ M., THEIN S. L., OLD J. M., LYNCH J. R., FALUSI A. G., WEATHERALL D.J. & CLEGG J. B., 1986. Evolutionary relationships o/human populations from an analysis of nuclear DNA polymorphisms. Nature, 319: 491-493.

WAINSCOAT J. S., HILL A. V. S. & CLEGG J. B., 1986. Reply to Giles and Ambrose. Nature, 322: 22. WILEY E. O., 1981. Phylogenetics. John Wiley and Sons, New York, Chichester. WOLPOFF M. H., Wu X. Z. & THORNE A. G., 1984. Modem Homo sapiens origins: a general theory of

hominid evolution involving the fossil evidence from East Asia In: (F. H. Smith & F. Spencer eds.). The origins of modern humans: a world survey of the fossil evidence, pp. 411-483. Alan R. Liss.

' Note added in proofs: a summary of the comments developed here was published in Nature (vol. 329: 111, 1987) followed by Cann, Stoneking and Wilson's answer (Nature, vol. 329: 111-112, 1987).

Received: 25 May 1987. Accepted: 25 July 1987.