A Reanalysis of the Ancient Mitochondrial DNA Sequences Recovered from Neandertal Bones

1359 Mol. Biol. E ol. 19(8):1359–1366. 2002 q 2002 by he Socie y o Molecula Biology and E olu ion. ISSN: 0737-4038 A Reanalysis o he Ancien Mi ochond ial DNA Sequences Reco e ed om Neande al Bones Gab iel Gu ie´ ez,* Diego Sa´nchez,†and An onio Ma ı´n* *Depa amen o de Gene´ ica, Uni e sidad de Se illa, Spain; and †Ins i u o de Biologı´a y Gene´ ica Molecula , Uni e sidad de Valladolid, Spain Recen epo s analyzing mi ochond ial DNA sequences om Neande al bones ha e claimed ha Neande als and mode n humans a e di e en species. The phylogene ic analyses ca ied ou in hese a icles did no ake in o accoun he high subs i u ion a e a ia ion among si es obse ed in he human mi ochond ial D-loop egion and also lack an es ima ion o he pa ame e s o he nucleo ide subs i u ion model. The sepa a e phylogene ic posi ion o Neande als is no suppo ed when hese ac o s a e conside ed. Ou analysis shows ha Neande al-Human and Human-Human pai wise dis ance dis ibu ions o e lap mo e han wha p e ious s udies sugges ed. We also show ha he mos ancien Neande al HVI egion is he mos di e gen when compa ed wi h mode n human sequences. Howe e , he opposi e would be expec ed i he sequence had no been modi ied since he dea h o he specimen. Such incong uence is discussed in he ligh o diagene ic modi ica ions in ancien Neande al DNA sequences. In oduc ion The classical iew eme ging om ana omical and a cheological s udies places Neande als as a di e en species om Homo sapiens. This is in ag eemen wi h he Ou -o -A ica hypo hesis (Cann, S oneking, and Wil- son 1987), which p edic s ha Neande als coexis ed wi hou ma ing wi h mode n humans who o igina ed in A ica om 100,000 o 200,000 yea s ago (S inge and And ews 1988). Ins ead, ecen ana omical and paleon- ological esea ch (Wolpo e al. 2001) suppo s he mul i egional hypo hesis, which p opounds ha some popula ions o a chaic Homo e ol ed in o mode n hu- man popula ions in many egions. Consequen ly, Nean- de als could ha e con ibu ed o he gene ic pool o p esen -day Eu opeans. S udies abou mode n human gene ic di e si y (Foley 1998; Jo de e al. 2000) assume ha Neande als we e no ela ed o mode n humans, al hough such as- sump ions ha e been ex ensi ely deba ed (No dbo g 1998; Hawks and Wolpo 2001; Rele h o d 2001). Fi e mi ochond ial DNA D-loops ha e been eco - e ed om Neande al bones (K ings e al. 1997; K ings e al. 1999; K ings e al. 2000; O chinniko e al. 2000). Hence o h, we will e e o hese esea ch g oups as he Neande al sequencing g oups (NSGs). The phylogene - ic analyses o hese sequences loca ed Neande al DNA a he base o mode n human di e si y ( ig. 2C), sug- ges ing ha he Neande al genes p obably did no con- ibu e o he mode n human gene ic pool. Howe e , he human D-loop egion shows an ex- eme a ia ion in he subs i u ion a e among si es (Ex- co ie and Yang 1999; Meye , Weiss, and on Haesele 1999) and a high amoun o pa allel mu a ions (Tamu a and Nei 1993) ha complica e phylogene ic econs uc- ions (Maddison, Ru olo and Swo o d 1992; Ingman e al. 2000). Mo eo e , om a phylogene ic poin o iew, he NSG nei he selec ed o he bes model o nucleo- Key wo ds: Neande al, ancien DNA, D-loop. Add ess o co espondence and ep in s: Gab iel Gu ie´ ez, De- pa amen o de Gene´ ica. Uni e sidad de Se illa, Apa ado 1095, 41080 Se illa, Spain. E-mail: [email p o ec ed]. ide subs i u ion no es ima ed he model pa ame e s ( a- ble 1). An abundan li e a u e (Yang, Goldman, and F i- day 1994; Swo o d e al. 1996; Huelsenbeck and Ran- nala 1997; B occhie i 2001; Posada and C andall 2001) emphasizes he need o an accu a e es ima ion o he nucleo ide subs i u ion model and pa ame e s, co ec ion o among-si e a e a ia ion, axa sampling, and selec- ion o he ou g oup o a ain a co ec ee econs uc ion. The pu pose o his a icle is o eanalyze he a ail- able Neande al DNA sequence da a. A e selec ing he app op ia e ex an human D-loop da a se s, we used he maximum-likelihood p inciple o model selec ion and pa ame e es ima ion o pe o m a no el phylogene ic econs uc ion. Da a and Me hods Sequences and Da a Se s We ha e used he MOUSE 1.0 da abase (Bu cha d , on Haesele , and Meye 1999), an aligned compila ion o m DNA con ol egions o p ima e species ha con- ained he Feldho e Neande al sequences. An ea lie e sion o his da abase (H Base) was used by he NSG. O he Neande al sequences we e e ie ed om GenBank ( able 1) and manually aligned o ou MOUSE-selec ed da a se s. We ca ied ou wo analyses: he i s analysis con- ce ns he HVI egion and includes h ee Neande al se- quences. The second analysis conce ns he HVI plus HVII egions including wo Neande al sequences. The HVII egion o he Mezmaiskaya specimen has no been eco e ed. In he i s analysis (HVI egion) we included hu- man en ies con aining a leas posi ions 16056 o 16378 o he Ande son e al. (1981) e e ence sequence ( ig. 1). This segmen co e s he minimal egion common o he h ee Neande al en ies. Human sequences, dupli- ca ed o ambiguous, we e dele ed. This selec ion en- de ed 1,905 human sequences (328 A icans, 471 Asians, 211 Aus alians-Oceanians, 475 Eu opeans, and 420 Ame icans). Th ee Pan paniscus and h ee Pan oglody es sequences we e added as ou g oup. Also, we Downloaded om h ps://academic.oup.com/mbe/a icle/19/8/1359/997686 by Uni e sidad de Se illa use on 23 Ma ch 2022 1360 Gu ie´ ez e al. Table 1 Repo ed Neande al Mi ochond ial DNA Sequences and Me hods Used o Phylogene ic Analysis Pape Da a Se S udied Sequence Reco - e ed Leng h Sequence Bounda ies a Geog aphic Region T ee Recons uc ion Me hod Model o nucleo ide Subs i u ion and P og am Used B anch Suppo K ings e al. 1997 ...... HVI 379 16023–16400 Feldho e , Ge - many Neighbo -join- ing Maximum-likelihood dis- ances (PHYLIP) Likelihood mapping (TREE-PUZZLE) F84 model HVI....... TS/TV a io 20 (no es i- ma ed) No gamma K ings e al. 1999 ...... HVII 345 57–398 Feldho e , Ge - many Neighbo -join- ing Maximum-likelihood dis- ances (PHYLIP) Likelihood mapping (TREE-PUZZLE) F84 model HVI 1 HVII .... TS/TV a io es ima ed wi h TREE-PUZZLE No gamma O chinniko e al. 2000 ...... HVI 345 16056–16399 Mezmaiskaya, No he n Caucasus Neighbo -join- ing, Maxi- mum pa si- mony Co ec ed dis ances Tamu a-Nei model wi h p og am de aul s Gamma wi h a shape al- ue 0.4 Boo s ap (PAUP*) HVI....... (PAUP*) K ings e al. 2000 ...... HVI 357 16023–16378 Vindija, C oa ia Neighbo -join- ing Maximum-likelihood dis- ances (PHYLIP) Likelihood mapping (TREE-PUZZLE) HVII 288 57–343 F84 model HVI 1 HVII .... TS/TV a io es ima ed wi h TREE-PUZZLE No gamma a Taking he sequence shown in Ande son e al. (1981) as a e e ence. F IG . 1.—Minimum anges equi ed o human HVI and HVII sequences o be selec ed. The minimum ange co e s he common egion o h ee HVI o wo HVII Neande al sequences. The maximum ange co e s he longes Neande al HVI o HVII sequence. The maximum ange se les he alignmen leng h. The numbe s co espond o he anges co e ed by he MOUSE da abase acco ding o he Ande son e al. (1981) e e ence sequence. disca ded alignmen columns whe e one sequence showed an inse ion, whe eas he es did no show any inse ion. The inal alignmen con ains 3 Neande al, 1,905 human, and 6 chimpanzee sequences. The ange used o analysis co e s posi ions 16023–16400, ha co espond o he longes Neande al sequence ( ig. 1). In he second analysis we included indi iduals whe e bo h HVI and HVII egions we e sequenced a leas om posi ions 16056–16378 and 57–343, espec i ely ( ig. 1). These anges co e he egion common o he Downloaded om h ps://academic.oup.com/mbe/a icle/19/8/1359/997686 by Uni e sidad de Se illa use on 23 Ma ch 2022 Neande al m DNA Phylogeny 1361 Table 2 Pa ame e Es ima ion o he Tamu a-Nei Model by he Maximum-Likelihood me hod D ATA S ET B ASE C OMPOSITION ACGT R ATE M ATRIX A ↔ CA ↔ GA ↔ TC ↔ GC ↔ TG ↔ T R ATE HETEROGENEITY P ARAMETERS Gamma Shape Pa ame e a P opo ion o In a iable Si es 1,914 HVI sequences 1,905 Humans..... 3 Neande als ..... 3Pan oglody es .. 3Pan paniscus .... 0.34 0.36 0.11 0.20 1 21.78 1 1 28.73 1 0.70 0.27 385 HVI 1 HVII sequences 377 Humans ...... 2 Neande als ..... 3Pan oglody es .. 3Pan paniscus .... 0.32 0.35 0.14 0.19 1 16.30 1 1 29.62 1 0.40 0.26 a Fou ca ego ies we e used o disc e e gamma dis ibu ion. wo Neande al sequences. Repea ed o ambiguous en- ies we e dele ed. A o al o 377 human en ies was selec ed (79 A icans, 59 Asians, 84 Aus alians-Ocean- ians, 123 Eu opeans, and 32 Ame icans). The inal alignmen con ains 2 Neande al, 377 human, and 6 chimpanzee sequences ha co e posi ions 16023– 16400 and 57–343. We dele ed alignmen columns whe e mos sequences showed a gap. Model Tes ing In o de o selec he bes model o nucleo ide sub- s i u ion i ing each sequence se , we ha e ollowed he p ocedu e desc ibed by Posada and C andall (1998), im- plemen ed in he MODELTEST p og am. This app oach i s ob ains a neighbo -joining ee (Sai ou and Nei 1987), using he Jukes-Can o model o nucleo ide sub- s i u ion (Jukes and Can o 1969) and hen uses a like- lihood a io es s a is ic o selec he bes model and es ima e i s pa ame e s, keeping he same ee opology. Pai wise Dis ances Unco ec ed and maximum-likelihood pai wise dis- ances ha e been compu ed using PAUP* 4.0 (Swo o d 1998). Maximum-likelihood dis ances we e calcula ed using he bes model i ing he da a. Phylogene ic Analysis and Boo s apping Phylogene ic analyses we e ca ied ou using PAUP* 4.0 (Swo o d 1998). Fo each da a se he p o- cedu e was: (1) he sequence o de was andomized, (2) he maximum-likelihood dis ances (using he selec ed model and he bes pa ame e s) we e compu ed, and (3) a neighbo -joining ee (Sai ou and Nei 1987) was con- s uc ed o 100 boo s ap eplica es (Felsens ein 1985). In e io b anch es s (Si niko a, Rzhe sky, and Nei 1995) we e ca ied ou wi h he p og am PHYLTEST 2.0 (Kuma 1996) using he mos complex model a ail- able in his p og am: he Kimu a wo-pa ame e model (Kimu a 1980) wi h gamma co ec ion. Resul s and Discussion Selec ion o Bes Model o Nucleo ide Subs i u ion, and Maximum-Likelihood Es ima ion o he Model Pa ame e s The bes model i ing he HVI and HVI 1 HVII se s is he Tamu a-Nei wi h a gamma dis ibu ion o a e he e ogenei y and a p opo ion o in a iable si es, ou o he 56 di e en models implemen ed in he MO- DELTEST p og am. Ac ually, his model was de eloped o he p ima e D-loop egion (Tamu a and Nei 1993). The ad an age o applying a co ec ion o a e a ia ion among si es in his egion has also been epo ed (Ta- mu a and Nei 1993; Wakeley 1993; Exco ie and Yang 1999; Meye , Weiss, and on Haesele 1999). I has been shown ha co ec ion o a e a ia ion among si es and conside a ion o in a iable si es imp o e phylogene ic econs uc ion in cases whe e a long b anch a ac ion (LBA) a i ac exis s (Philippe and Lau en 1998). As we will show la e , chimpanzee sequences can a ac he Neande al ones. The pa ame e s es ima ed in ou wo k ( able 2) a e simila o hose ob ained by o he au ho s (Exco ie and Yang 1999; Meye , Weiss, and on Haesele 1999). The e is a s ong di e ence be ween ansi ions and ans e sions, and a low alue o he gamma dis ibu- ion shape pa ame e , which indica es a s ong he e o- genei y in he subs i u ion a e among si es. The alue o he shape pa ame e es ima ed he e is g ea e han o he es ima ions (Exco ie and Yang 1999; Meye , Weiss, and on Haesele 1999) because we ha e in o- duced a pa ame e o he p opo ion o in a iable si es. Phylogene ic Recons uc ion and Boo s ap Schema ic ep esen a ions o he boo s ap ee o- pologies de i ed om ou HVI and HVI 1 HVII da a se s a e shown in igu e 2Aand B. The de ailed opol- ogies, including 1,914 and 385 sequences, espec i ely, a e no shown he e, bu a e a ailable upon eques . Boo s ap alues we e no signi ican o mos nodes. Al hough he h ee Neande al sequences clus e ed o- ge he , we ob ained no suppo o a b anch sepa a ing Downloaded om h ps://academic.oup.com/mbe/a icle/19/8/1359/997686 by Uni e sidad de Se illa use on 23 Ma ch 2022 1362 Gu ie´ ez e al. F IG . 2.—Phylogene ic ela ionships o Neande als. Chimpanzee sequences a e used as ou g oup. (A) The boo s ap majo i y ule consensus ee de i ed om he HVI da a se was ob ained when conside ing he bes - i model o nucleo ide subs i u ion and bes es ima ion o model pa ame e s (see able 2). (B) The same as in (A) o he HVI 1 HVII da a se . In (C) we show a ep esen a ion o he ees and b anch suppo ob ained by, K ings e al. (1997), K ings e al. (1999), K ings e al. (2000), and O chinniko e al. (2000). NSG, Neande al sequencing g oups. he Neande al clus e om he human sequences. Ou HVI ee ( ig. 2A) is simila o he NSG one, bu b anch suppo is no signi ican . The HVI 1 HVII ee ( ig. 2B) places 10 A icans sequences as ou g oup o Nean- de als and o he humans, bu again, he boo s ap alue is no signi ican . The e o e, ou esul s sugges ha he Neande al sequences canno be conside ed an ou g oup o mode n humans. This esul is in con as o he ee ob ained by he NSG, whe e he Neande al sequences appea basal o mode n humans when using chimpanzee sequences as ou g oup ( ig. 2C). Gi en ha nei he he boo s ap me hod (Felsen- s ein 1985) no he likelihood mapping me hod used by K ings e al. (1997, 1999, 2000) ha e s a is ical mean- ing, we pe o med he in e io b anch es . I has been shown ha he con idence alue o his es is he com- plemen o he P alues in he s anda d s a is ical es (Si niko a, Rzhe sky, and Nei 1995). We di ided ou da a se s in o ou clus e s: (1) chimpanzees, (2) Nean- de als, (3) ex an human sequences loca ed a he base o he HVI 1 HVII ee ( ig. 2B) ha co espond o 10 !Kung en ies o he MOUSE da abase, and (4) he es o humans. We hen de e mined whe he he leng hs o he in e io b anches o he h ee possible ees o he ou sequence clus e s a e signi ican ly di e en om 0. The esul o his es is no signi ican o ei he he HVI o HVI 1 HVII egions. Thus, like he boo s ap analyses, he in e io b anch es sugges s ha a poly- omy is he bes ep esen a ion o he e olu iona y ee ela ing Neande al and ex an humans. Pai wise Dis ances The NSG epo ed ha he pai wise compa isons be ween he Neande al and human sequences demon- s a e ha Neande als a e ou side o mode n human D- loop a iabili y. In pa icula , K ings e al. (1997) s a ed ha ‘a o al o 0.002% o he pai wise compa isons be- ween human m DNA sequences we e la ge han he smalles di e ence be ween he Neande al and he hu- mans.’ We hink ha his poin me i s u he analysis. The cu en da abase is biased because o he o e ep- esen a ion o some popula ions and he unde ep esen- a ion o o he s. Fo ins ance, he MOUSE da abase con- ains 6,012 en ies o he HVI egion, bu 31% o he en ies belong o only 20 popula ions ou o 206 popu- la ions ep esen ed (10% o he o al o popula ions). The ex eme cases a e 306 Ko eans, 126 Yaps, 120 Cay- apa Ame indians, 119 Mandeka, 115 Palau, and 100 whi e B i ish. The e a e also 1,417 en ies o unde e - mined popula ion (40% o hem a e om No h Ame ica and 23% Eu opean, bu only 9% a e om A ica). Thus, A ican popula ions con aining he mos ancien lineages and he highes a ia ion a e unde ep esen ed in he da abase. Because o he da abase o e ep esen a ion o some human popula ions, he dis ibu ion o pai wise dis anc- es is biased. A la ge pa o pai wise compa isons a e made be ween indi iduals belonging o he same pop- ula ion. Likewise, i is expec ed ha mos indi iduals om a single popula ion will show simila dis ances o a gi en ou g oup (Neande al, in his case). To o e come his p oblem, we conside ed ano he sample o he hu- man a ia ion. We i s so ed he HVI sequences in ou da a se acco ding o i s unco ec ed dis ance o he e - e ence sequence (Ande son e al. 1981). Then we g ouped hem in o 171 classes, con aining equidis an sequences (conside ing ou decimals), and chose one sequence a andom om each class. The compu a ion Downloaded om h ps://academic.oup.com/mbe/a icle/19/8/1359/997686 by Uni e sidad de Se illa use on 23 Ma ch 2022 Neande al m DNA Phylogeny 1363 F IG . 3.—The h ee possible qua e s be ween he HVI sequence om he Feldho e Neande al, a chimpanzee, and wo humans. The human A is one o he mos di e gen sequences wi h espec o he e e ence sequence (Ande son e al. 1981), and he human B is one o he less di e gen . Qua e s we e e alua ed wi h he p og am TREE-PUZZLE unde he F84 model ( he ansi ion- ans e sion a io was 20). Thei espec i e log-likelihoods a e shown below hem. o pai wise dis ances be ween he 171 andomly selec ed sequences and he Neande als ende ed 1.6% o human- human compa isons la ge han he smalles di e ence be ween Neande als and humans. Likewise, 27% o he compa isons a e lowe han he la ges human-human di e ence. This esul sugges s ha Neande als se- quences a e no so di e en om hose o ex an hu- mans, in con as o he NSG claims. Final Conside a ions A main conclusion can be ex ac ed om ou anal- yses: he phylogene ic posi ion o he ancien DNA se- quences eco e ed om Neande al bones is sensi i e o he phylogene ic me hods employed. I depends on he model o nucleo ide subs i u ion, he b anch suppo me hod, and he se o da a used. Adcock e al. (2001) eco e ed HVI sequences o a chaic human bones om Aus alia, and hei phylogene ic analysis showed ha wo o he specimens we e ou g oups e en o he mos ancien A ican lineages. They concluded ha his is an e idence suppo ing he mul i egional hypo hesis. How- e e , a second analysis ca ied ou by Coope e al. (2001) ha ook in o accoun he he e ogenei y o a es be ween si es and a la ge sample o mode n humans, showed ha bo h HVI sequences a e loca ed among ex- an humans. This case illus a es he in luence o he nucleo ide subs i u ion model on he phylogene ic e- cons uc ion o he human D-loop egion. The NSG s udies used poo pa ame e models o nucleo ide subs i u ion o hei analyses, whe eas we op ed o complex (pa ame e ich) models ollowing he likelihood a io es . Bo h al e na i es ha e p os and cons. Yang (1997) and Posada and C andall (2001) en- cou age he use o he bes - i model, al hough hey e- po ed some examples whe e simple models can eco e he ue phylogene ic ee be e han complex models. Se e al au ho s highligh he impo ance o using he bes nucleo ide subs i u ion model o a gi en da a se (Huelsenbeck and Hillis 1993; Sulli an and Swo o d 1997; Cunningham, Zhu, and Hillis 1998). The same conside a ions apply o he choice o he b anch suppo me hod. The boo s ap me hod has some ca ea s (Cum- mings, O o, and Wakeley 1995), and he qua e puz- zling (likelihood mapping) has also been c i icized o o e es ima ing b anch suppo alues (Cao, Adachi, and Hasegawa 1998). Secondly, we belie e ha he likelihood mapping alues suppo ing Neande als as a di e en species migh be a i ac ually inc eased. To illus a e his poin , we ook wo human HVI sequences showing maximal di e gence be ween hem, a chimpanzee, and he Feld- ho e Neande al. The h ee possible qua e s we e e al- ua ed wi h he p og am TREE-PUZZLE (S imme and on Haesele 1996) wi h he same model and pa ame e used by K ings e al. (1997). I can be seen ( ig. 3) ha he chimpanzee b anch is e y long, and ha he Ne- ande al b anch is longe han he wo humans, opening he possibili y o a LBA a i ac . I has been epo ed ha qua e me hods a e e y p one o LBA (Huelsen- beck 1998). Thus, he bes -sco ed opology is usually he one joining long b anches in exhaus i e qua e e al- ua ions. Ano he issue ela es o he signi icance o he log-likelihood alues. In he ees d awn in igu e 3, he bes sco e is o he ee ha joins Neande al and chim- panzee sequences. Howe e , he log-likelihoods a e e y simila be ween he al e na i e opologies. The Kishino- Hasegawa es (Kishino and Hasegawa 1989) yielded no signi ican di e ences be ween hem a he 5% le el. This example illus a es ha likelihood mapping can se- lec as he bes opology one ha is no signi ican ly be e han he o he wo. A hi d a gumen agains he use o likelihood mapping in his s udy has o do wi h he da abase popula ion bias men ioned ea lie . I is no ad isable o pe o m likelihood mapping wi h a da abase Downloaded om h ps://academic.oup.com/mbe/a icle/19/8/1359/997686 by Uni e sidad de Se illa use on 23 Ma ch 2022 1364 Gu ie´ ez e al. F IG . 4.—Compa ison be ween ancien DNA sequences o wo Neande al specimens and ex an humans. Neande al specimen A ( he mos ancien ) would be expec ed o show a mean dis ance o humans sho e han he mos ecen one (Neande al B) i DNA had su e ed no damage. Table 3 Mean Dis ances Be ween Neande al and 1905 Mode n Human HVI Sequences Neande al Sequence Da e a p Dis ance Maximum- Likelihood Dis ance b Feldho e ........ Mezmaiskaya .... Vindija.......... 40,000–100,000 29,195 42,000 0.0801 0.0689 0.0677 0.1050 0.0898 0.0869 a Yea s be o e p esen . b Unde he bes model, Tamu a-Nei wi h gamma o a e he e ogenei y and he model pa ame e s shown in able 2. whe e mos human sequences a e e y simila among hemsel es. O chinniko e al. (2000) also ob ained a high boo s ap alue (95%) o he Neande al clade, seg e- ga ing hem om mode n humans. When we used he same phylogene ic p ocedu es, i.e., co ec ed Tamu a- Nei dis ances wi h a gamma dis ibu ion o 0.4, we ob- ained a lowe boo s ap alue (62%). An explana ion o his disc epancy could be ha O chiniko e al. did no selec human sequences by leng h and did no elim- ina e sequences wi h ambigui ies. Missing da a a e a common nuisance ac o in phylogene ic analysis (Wiens 1998). Al hough we ob ained a lowe boo s ap alue, we hink ha 62% is s ill a high boo s ap alue compa ed wi h he ones ob ained when maximum-like- lihood dis ances a e used ( ig. 2Aand B). I has been shown ha maximum-likelihood dis ances ha e some ad an ages o e co ec ed dis ances, educing sampling a iances o he ansi ion- ans e sion a io es ima ions (Swo o d e al. 1996). Meye , Weiss, and on Haesele (1999) es ima ed a ansi ion- ans e sion a io o 15.55 o he HVI egion. This poin s o an unde es ima ion by O chinniko e al. (2000) o he numbe o subs i- u ions along he ee b anches. The low boo s ap alues ob ained in ou wo k in- dica e he lack o esolu ion o he HVI and HVII se- quences o de e mine he phylogene ic posi ion o Nean- de als because no ee opology is a o ed. In pa , his is caused by sho sequences wi h a high p opo ion o in a iable si es. Ingman e al. (2000) no ed ha he D- loop egion, in spi e o i s wide use o human phylo- gene ic analysis, is no an app op ia e egion o he mi- ochond ial genome o answe ing phylogene ic ques ions. Ano he in e es ing issue is he possibili y ha he Neande al sequences we e a i ac ual. Calda a o and Gabow (2000) no iced ha some o he nucleo ide sub- s i u ions ound in he Feldho e Neande al sequence ma ched hose ound in o he human ancien HVI se- quences. They a ibu ed hese simila i ies o diagene ic changes ha occu ed in ossil DNA. Hansen e al. (2001) s udied di e en clones o ancien DNA ampli- ica ions and concluded ha PCR can in oduce e o s caused by he miscoding lesions o ancien DNA. In he absence o diagene ic changes, he mean dis ance be- ween an ancien sequence and an ex an one ough o be la ge han he dis ance be ween a mo e ancien se- quence and he ex an one. So, he mo e ancien a se- quence, he sho e i s dis ance o he ex an sequences ( ig. 4). We es ed his p edic ion by compu ing he a - e age dis ances be ween each Neande al HVI sequence and ou da a se o ex an HVI human sequences ( able 3). I can be seen ha he mos ecen specimen (Mez- maiskaya) shows a sho e dis ance han he oldes one (Feldho e ). This esul can be explained by a di e en nucleo ide subs i u ion a e among Neande al lineages (popula ions) o by miscoding DNA lesions in Nean- de al ossils. Had DNA damage inc eased he di e - ences be ween Neande al and mode n humans, Nean- de als would be mo e akin o mode n humans han wha ecen claims sugges . Acknowledgmen s We hank G. Gu ie´ ez-P ie o and wo anonymous e e ees o help ul commen s on he manusc ip . This Downloaded om h ps://academic.oup.com/mbe/a icle/19/8/1359/997686 by Uni e sidad de Se illa use on 23 Ma ch 2022 Neande al m DNA Phylogeny 1365 wo k was suppo ed in pa by he DGICYT BIO-1999- 065-C02-02 o he Spanish Go e nmen . LITERATURE CITED A DCOCK , G. J., E. S. D ENNIS ,S.E ASTEAL ,G.A.H UTTLEY ,L. S. J ERMIIN ,W.J.P EACOCK , and A. T HORNE . 2001. Mi o- chond ial DNA sequences in ancien Aus alians: implica- ions o mode n human o igins. P oc. Na l. Acad. Sci. USA 98:537–542. A NDERSON , S., A. T. B ANKIER ,B.G.B ARRELL e al. (14 co- au ho s). 1981. Sequence and o ganiza ion o he human mi ochond ial genome. Na u e 290:457–465. B ROCCHIERI , L. 2001. Phylogene ic in e ences om molecula sequences: e iew and c i ique. Theo . Popul. Biol. 59:27– 40. B URCHARDT , F., A. VON H AESELER , and S. M EYER . 1999. H Base: compila ion o m DNA con ol egion sequences om p ima es. Nucleic Acids Res. 27:138–142. C ALDARARO , N., and S. G ABOW . 2000. Mi ohond ial DNA analysis and he place o Neande als in Homo. Ancien Biomol. 3:135–158. C ANN , R. L., M. S TONEKING , and A. C. W ILSON . 1987. Mi o- chond ial DNA and human e olu ion. Na u e 325:31–36. C AO , Y., J. A DACHI , and M. H ASEGAWA . 1998. Commen on he qua e me hod o inding maximum-likelihood ee o- pologies. Mol. Biol. E ol. 15:87–89. C OOPER , A., A. R AMBAUT ,V.M ACAULAY ,E.W ILLERSLEV ,A. J. H ANSEN , and C. S TRINGER . 2001. Human o igins and an- cien human DNA. Science 292:1655–1656. C UMMINGS , M. P., S. P. O TTO , and J. W AKELEY . 1995. Sam- pling p ope ies o DNA sequence da a in phylogene ic analysis. Mol. Biol. E ol. 12:814–822. C UNNINGHAM , C. W., H. Z HU , and D. M. H ILLIS . 1998. Bes - i maximum-likelihood models o phylogene ic in e ence: empi ical es s wi h known phylogenies. E olu ion 52:978– 987. E XCOFFIER , L., and Z. Y ANG . 1999. Subs i u ion a e a ia ion among si es in mi ochond ial hype a ible egion I o hu- mans and chimpanzees. Mol. Biol. E ol. 16:1357–1368. F ELSENSTEIN , J. 1985. Con idence limi s on phylogenies: an app oach using he boo s ap. E olu ion 39:783–791. F OLEY , R. 1998. The con ex o human gene ic e olu ion. Ge- nome Res. 8:339–347. H ANSEN , A. J., E. W ILLERSLEV ,C.W IUF ,T.M OURIER , and P. A RCTANDER . 2001. S a is ical e idence o miscoding le- sions in ancien DNA empla es. Mol. Biol. E ol. 18:262– 265. H AWKS , J., and M. W OLPOFF . 2001. Paleoan h opology and he popula ion gene ics o ancien genes. Am. J. Phys. An h o- pol. 114:269–272. H UELSENBECK , J. P. 1998. Sys ema ic bias in phylogene ic analysis: is he S epsip e a p oblem sol ed? Sys . Biol. 47: 519–537. H UELSENBECK , J. P., and D. M. H ILLIS . 1993. Success o phy- logene ic me hods in he ou - axon case. Sys . Biol. 42: 247–264. H UELSENBECK , J. P., and B. R ANNALA . 1997. Phylogene ic me hods come o age: es ing hypo hesis in an e olu iona y con ex . Science 276:227–232. I NGMAN , M., H. K AESSMANN ,S.P A ¨A ¨BO , and U. G YLLENSTEN . 2000. Mi ochond ial genome a ia ion and he o igin o mode n humans. Na u e 408:708–713. J ORDE , L. B., W. S. W ATKINS ,M.J.B AMSHAD ,M.E.D IXON , C. E. R ICKER ,M.T.S EIELTAD , and M. A. B ATZER . 2000. The dis ibu ion o human gene ic di e si y: a compa ison o mi ochond ial, au osomal, and Y-ch omosome da a. Am. J. Hum. Gene . 66:979–988. J UKES , T. H., and C. R. C ANTOR . 1969. E olu ion o p o ein molecules. Pp. 21–132 in H. N. M UNRO , ed. Mammalian p o ein me abolism. Academic P ess, New Yo k. K IMURA , M. 1980. A simple me hod o es ima ing e olu ion- a y a es o base subs i u ions h ough compa a i e s udies o nucleo ide sequences. J. Mol. E ol. 16:111–120. K ISHINO , H., and M. H ASEGAWA . 1989. E alua ion o he max- imum likelihood es ima ion o he e olu iona y ee opol- ogies om DNA sequence da a, and he b anching o de in Hominoidea. J. Mol. E ol. 29:170–179. K RINGS , M., C. C APELLI ,F.T SCHENTSCHER ,H.G EISERT ,S. M EYER ,A. VON H AESELER ,K.G ROSSSCHMIDT ,G.P OS - SNERT ,M.P AUNOVIC , and S. P A ¨A ¨BO . 2000. A iew o Ne- ande al gene ic di e si y. Na . Gene . 26:144–145. K RINGS , M., H. G EISERT ,R.W.S CHMITZ ,H.K RAINITZKI , and S. P A ¨A ¨BO . 1999. DNA sequence o he mi ochond ial hy- pe a iable egion II om he Neande al ype specimen. P oc. Na l. Acad. Sci. USA 96:5581–5585. K RINGS , M., A. S TONE ,R.W.S CHMITZ ,H.K RAINITZKI ,M. S TONEKING , and S. P A ¨A ¨BO . 1997. Neande al DNA sequenc- es and he o igin o mode n humans. Cell 90:19–30. K UMAR , S. 1996. PHYLTEST: phylogene ic hypo hesis es ing so wa e. Ve sion 2.0. Pennsyl ania S a e Uni e si y, Uni- e si y Pa k, Pa. M ADDISON , D. R., M. R UVOLO , and D. L. S WOFFORD . 1992. Geog aphic o igins o human mi ochond ial DNA: phylo- gene ic e idence om con ol egion sequences. Sys . Biol. 41:111–124. M EYER , S., G. W EISS , and A. VON H AESELER . 1999. Pa e n o nucleo ide subs i u ion and a e he e ogenei y in he hype - a iable egions I and II o human m DNA. Gene ics 152: 1103–1110. N ORDBORG , M. 1998. On he p obabili y o Neande hal an- ces y. Am. J. Hum. Gene . 63:1237–1240. O VCHINNIKOV , I. V., A. G O ¨THERSTRO ¨M ,G.O.R OMANOVA ,V. M. K HARITONOV ,K.L IDE ´N , and W. G OODWIN . 2000. Mo- lecula analysis o Neande hal DNA om he no he n Cau- casus. Na u e 404:490–493. P HILIPPE , H., and J. L AURENT . 1998. How good a e deep phy- logene ic ees? Cu . Opin. Gene . De . 8:616–623. P OSADA , D., and K. A. C RANDALL . 1998. MODELTEST: es - ing he model o DNA subs i u ion. Bioin o ma ics 14:817– 818. ———. 2001. Simple (w ong) models o complex ees: a case om e o i idae. Mol. Biol. E ol. 18:271–275. R ELETHFORD , J. H. 2001. Absence o egional a ini ies o Ne- ande al DNA wi h li ing humans does no ejec mul i e- gional e olu ion. Am. J. Phys. An h opol. 115:95–98. S AITOU , N., and M. N EI . 1987. The neighbo -joining me hod: a new me hod o econs uc ing phylogene ic ees. Mol. Biol. E ol. 4:406–425. S ITNIKOVA , T., A. R ZHETSKY , and M. N EI . 1995. In e io - b anch and boo s ap es s o phylogene ic ees. Mol. E ol. Biol. 12:319–333. S TRIMMER , K., and A. VON H AESELER . 1996. Qua e puzzling: a qua e maximum likelihood me hod o econs uc ing ee opologies. Mol. Biol. E ol. 13:964–969. S TRINGER , C. B., and P. A NDREWS . 1988. Gene ic and ossil e idence o he o igin o mode n humans. Science 239: 1263–1268. S ULLIVAN , J., and D. L. S WOFFORD . 1997. A e guinea pigs oden s? The impo ance o adequa e models in molecula phylogene ics. J. Mamm. E ol. 4:77–86. Downloaded om h ps://academic.oup.com/mbe/a icle/19/8/1359/997686 by Uni e sidad de Se illa use on 23 Ma ch 2022 1366 Gu ie´ ez e al. S WOFFORD , D. L. 1998. PAUP*: phylogene ic analysis using pa simony (*and o he me hods). Ve sion 4. Sinaue Asso- cia es, Sunde land, Mass. S WOFFORD , D. L., G. J. O LSEN ,P.J.W ADDELL , and D. M. H ILLIS . 1996. Phylogene ic in e ence. Pp. 407–514 in D. M. H ILLIS ,C.M ORITZ , and B. K. M ABLE , eds. Molecula sys- ema ics. Sinaue , Sunde land, Mass. T AMURA , K., and M. N EI . 1993. Es ima ion o he numbe o nucleo ide subs i u ions in he con ol egion o mi ochon- d ial DNA in humans and chimpanzees. Mol. Biol. E ol. 10:512–526. W AKELEY , J. 1993. Subs i u ion a e a ia ion among si es in hype a iable egion 1 o human mi ochond ial DNA. J. Mol. E ol. 37:613–623. W IENS , J. J. 1998. Combining da a se s wi h di e en phylo- gene ic his o ies. Sys . Biol. 47:568–581. W OLPOFF , M. H., J. H AWKS ,D.W.F RAYER , and K. H UNLEY . 2001. Mode n human ances y a he pe iphe ies: a es o he eplacemen heo y. Science 291:293–297. Y ANG , Z. 1997. How o en do w ong models p oduce be e phylogenies? Mol. Biol. E ol. 14:105–108. Y ANG , Z., N. G OLDMAN , and A. F RIDAY . 1994. Compa ison o models o nucleo ide subs i u ion used in maximum-like- lihood phylogene ic es ima ion. 1994. Mol. Biol. E ol. 11: 316–324. M ANOLO G OUY , e iewing edi o Accep ed Ap il 15, 2002 Downloaded om h ps://academic.oup.com/mbe/a icle/19/8/1359/997686 by Uni e sidad de Se illa use on 23 Ma ch 2022