Supplementary information for Pierson et al (2006) Deciphering Past Human Population Movements in Oceania: Provably Optimal Trees of 127 mtdna Genomes

Transcription

1 Supplementary information for Pierson et al (2006) Deciphering Past Human Population Movements in Oceania: Provably Optimal Trees of 127 mtdna Genomes Supplementary Table 1: Pacific dataset details Haplogroup file names and details Supplementary Figure 1: B5a haplogroup minimal tree and branch-labelled Supplementary Figure 2: M7bc haplogroup with M22 consensus network and branch-labelled Supplementary Figure 3: M27 and M28 haplogroup minimal tree and branchlabelled Supplementary Figure 4: P haplogroup with R21 minimal tree and branchlabelled Supplementary Figure 5: Q haplogroup with M29 consensus network and branch-labelled

2 Supplementary Table 1: Pacific dataset details Location n Accession numbers Source Australia 26 AF AF AY AY DQ DQ Ingman et al 2000 Kivisild et al 2006 Island Southeast Asia Philippines Malaysia Nicobar Islands Andaman Islands AF AY AY AY AY AY AY AY Maca-Meyer et al 2001 Macaulay et al 2005 Thangaraj et al 2005 Thangaraj et al 2005 Taiwan 14 AY AY AJ AJ DQ DQ Trejaut et al 2005 Near Oceania New Guinea Bismarck Archipelago Bougainville Melanesian Trobriand Islands AF AF AY AY DQ DQ AY AY DQ DQ137404, DQ DQ AY AY DQ137405, DQ DQ DQ DQ DQ DQ Ingman et al 2000 Kivisild et al 2006 Friedlaender et al 2005 Merriwether et al 2005 Macaulay et al 2005 Merriwether et al 2005 Kivisild et al 2006 Remote Oceania Samoa Tonga Cook Islands Vanuatu Kapingamarangi Atoll Marshall Islands AF AY AY DQ AY DQ AY AY DQ DQ DQ DQ DQ DQ DQ37287 Ingman et al 2000 Total 136

3 HAPLOGROUP FILE NAMES AND DETAILS PwithR21.nex no. of taxa=24 as it includes the rcrs and two coding-only sequences aligned DQ Australia R12, and DQ Melanesian P2 which were not included in the parsimony analysis. QwithM29.nex no. of taxa=27; includes rcrs and 6 coding-only sequences not included in the analysis DQ DQ112887, DQ112895, DQ112896, DQ B4a.nex no. of taxa=48, includes rcrs. B5a.nex no. of taxa=12, includes rcrs. M7bcM22.nex no. of taxa=47, includes rcrs. M27M28.nex no. of taxa=15, includes rcrs. These files, and the Oceanic dataset are available from Dataset No. taxa analysed Pars. inform. chars. Pars. score MMS max. No. of trees found B4a B5a M27M M7bcM PwithR Q with M NOTES ON ANALYSIS OF INSERTION/DELETION SITES Parsimony informative sites from the entire mtdna - coding and control region - were included in the haplogroup analyses, however gapped sites were treated as missing data, and excluded. In practise this means extensions and deletions of the C-stretches at n309 and n315 relative to the rcrs did not contribute to the analysis, and these are not shown on the branch-labelled trees in Figure 3 and Supp. Figures 1-5. At the short CA microsatellite region from n514-n523 the rcrs has (CA)x5; within the datasets there was variation of minus one, and plus one or two of these repeats relative to the rcrs. Where these extensions and contractions occurred it results in a gapped alignment. It should be noted that while these changes are shown on the branch labelled reconstructions they did not contribute to the analyses. At the variable C-stretch region in HVRI, between n16184-n16193, extensions in the C-stretch resulting in reduction in the flanking AAAA sequence were included in the parsimony analysis (see alignments, available from these are in nexus format, and have been character labelled for the rcrs sequence). Extensions beyond n16193 resulted in gaps in the alignments and thus were not included, and are not reconstructed on the labelled trees. Other deletions are marked on the labelled trees; again these were not included in the parsimony analysis but have been reconstructed to provide a complete listing of SNPs in the individual sequence (with the exception of n309c, n315c and 16193C extensions/deletions). The nine-base pair deletion present in several sequences at bases n8280-n8288 was artificially encoded in the parsimony analyses by adding a guanine at 8280 to the sequences in which it occurs, creating a transition rather than a gap at this character.

4 DQ Taiwan AY Nicobar Is. AY Nicobar Is. AY Nicobar Is. AY Nicobar Is. AY China APOO8451 Japan APOO8781 Japan APOO8263 Japan APOO8552 Japan a B5a b B R/N 9bpdel 16183CtvA(1) B5 709A 8584A 9950C 10398G 16140C B4 AY DQ B5a 210G(1) 3537G 15235G 16266AtvC(1) B5b AY AY AY AY AP AP AP AP T 11881T 13145A 13395G 93G 9962A 11149A 11151T 14149T 3540C 5773A 10754G 14989T 210A(2) 515G 6392C 16497G 8557A 8614C 15046G 15301A 16266G(2) 2550TtvA 12705T 13104G 13887G 14158T 16129A 16183A(2) 16187T 12234G 146C 5021C Supplementary Figure 1: B5a haplogroup minimal tree and branch-labelled a) The single most parsimonious minimal tree found, from 10 B5a individuals with L3 outgroup (entire mtdna sequence, no. of parsimony informative chars=17, parsimony score=21). Sequences from this study are shown in bold type. b) A base-labelled reconstructed from the minimal tree. Length variations in the poly-c region from nt303-nt315 are not shown. Other gapped sites in the alignment (for example at the short 'CA' microsatellite region from n514-n523) were not included in the analysis, but are shown on the labelled tree. Substitutions are transitions to the base shown, unless marked 'tvn' where N is the base in the rcrs. Synonymous transitions are shown in red type, and sites which change more than once in a lineage in the labelled are followed by the number of the change in brackets. When a substitution results in the same nucleotide as in the rcrs this is shown in italics. The polymorphisms relative to the rcrs at the N/R vertex are: 73G, 263G, 750G, 1438G, 2706G, 3106del, 4769G, 7028T, 8860G, 11719A, 14766T, 15326G.

5 AY Taiwan AP Japan AY Malaysia AY China AY China AY Taiwan AP Japan AY China AP Japan AP Japan AP Japan DQ Micronesia DQ Taiwan AY Mongolia AF Philippines a M22 AY M7bc AP AP M7c M7b AY AP AY AP AY AY b M AP AY AY M T DQ AF C 1119C 4639C 5351G 6071C 6776C 8158G 9316C 14142AtvC 15236G 16093C 16184T 16290T 16304C 3010A 4048A 4164G 5178AtvC 5351G 5460A 6261A 8414T 9128C 11696A 14668T M7b 4048A 4164G 5351G 5460A 6680C 7684C 7853A 12405T 16129A(1) 16297C 199C 6455T M7bc 9824C DQ M7c 146C(1) 4850T 5442C 11665T(1) 12091C 16295T(1) 30 individuals; 1 from China, 29 from Japan 150T 523insCA 709A 7175C 9103C 15040T 15912T 16192T 16224C 1664A 4454C 6351C 9468G 10497T 12121C 14115T 16086C 16324C 15244G 16309G 16509C 408AtvT 4216C 15849T 15857T 16176T 16192T 16526A 150T 12811C 3483A 7702A 9681G 16129G(2) 16301T 10345C 16298C 29 Japan, 1 China 1375T 1625G 7073G 7844G 8292A 12501A 12561A 373G 4452C 7960G 11389T 11914A 15787C 15924G 3882A 16319A 6053T 7521A 7961C 9017C 12804C 13269G 14281GtvC 14755TtvA 8718G 16294T 13299G 16295C(2) 146T(2) 513A 6503G 8059T 11152GtvT 11665C(2) 11719G 11946T 15236G 146AtvT(2) 955insC 5715G 6236T 8078A 16295C(2) 3606G 9bpdel 16129A 12579G Supplementary Figure 2: M7bc haplogroup with M22 consensus network and branch-labelled a) The consensus network of most parsimonious optimal trees found, from 44 M7bc individuals with the single M22 sequence and outgroup (entire mtdna sequence, no. of parsimony informative chars=56, parsimony score=93). Sequences from this study are shown in bold type. b) A base-labelled reconstructed from the consensus network; 30 M7b sequences are not shown in detail. Length variations in the poly-c region from nt303-nt315 are not shown. Other gapped sites in the alignment (for example at the short 'CA' microsatellite region from n514-n523) were not included in the analysis, but are shown on the labelled tree. Substitutions are transitions to the base shown, unless marked 'tvn' where N is the base in the rcrs. Synonymous transitions are shown in red type, and sites which change more than once in a lineage in the labelled are followed by the number of the change in brackets. When a substitution results in the same nucleotide as in the rcrs this is shown in italics. The polymorphisms relative to the rcrs at the M vertex are: 73G, 263G, 489C, 750G, 1438G, 2706G, 3106del, 4769G, 7028T, 8701G, 8860G, 9540C, 10398G, 10400T, 10873C, 11719A, 12705T, 14766T, 14783C, 15043A,15301A, 15326G, 16223T. The analysis suggests the APO8278 and AF sequences may contain errors.

6 M27 M28 DQ M27b M28a M28b DQ M DQ DQ DQ M27c M27a DQ DQ DQ A(1) (1) DQ DQ DQ DQ DQ M27b 14109T M27 64T 199C 236C 656C 2315G 3204T 3591A 3693A 4342G 4775G 5788C 5892C 10166C 10493C 14137T 14370G 14674C 16145A 16209C 16299G 16390A 5375T 9201T 12358G M27a 150T 195C 228A 234G 537T 739T 980C 2404C 5177A 5585A 7337A 8406T 9389G 10042G 11020G 11554GtvC 15451T 16048A 16077TtvA 16136C 16172C 16183CtvA 16311C 16320T (1) 1719G(2) M27c 146C 186T 1692G 3397G 8503C 9033G 9127G 11959G 12370T 14870G 16301T 16304C 16519T(2) 151T 204T(2) 207A 14894C 2951G 10752T 16320T M28a 234G 523insCA 1185T 3278C 3651G 14577C 16429T 10658G 16086C 16129A 523insCACA 7394G 16366T 16519T(2) 6962A 6858G 10192T 15530C 16311C 16468T(2) M28 195C 6281G 6374C 10245C 15067C 16148T 16468C(1) M28b 203A 508G 539C 574delA 1462A 3535C 4541A 6815C 9731AtvC 9944C 10670T 15739T 16093C 94A 12172G 16318TtvA 2135G 2280T 3808G 5772A 5951G 12373G 12507G 13754T 16291T 1503A 14730T 8376C DQ New Britain DQ Bougainville DQ Bougainville DQ Bougainville DQ New Britain DQ New Britain DQ New Ireland DQ New Britain DQ New Britain DQ Vanuatu DQ Vanuatu DQ New Britain DQ New Britain Supplementary Figure 3: M27 and M28 haplogroup minimal tree and branch-labelled a) The single most parsimonious minimal tree, from 13 M27 and M28 individuals and outgroup (entire mtdna sequence, no. of parsimony informative chars=82, parsimony score=95). Sequences from this study are shown in bold type. b) A base-labelled reconstructed from the minimal tree showing basal links at nt1719a and nt. Length variations in the poly-c region from nt303-nt315 are not shown. Other gapped sites in the alignment (for example at the short 'CA' microsatellite region from n514-n523) were not included in the analysis, but are shown on the labelled tree. Substitutions are transitions to the base shown, unless marked 'tvn' where N is the base in the rcrs. Synonymous transitions are shown in red type, and sites which change more than once in a lineage in the labelled are followed by the number of the change in brackets. When a substitution results in the same nucleotide as in the rcrs this is shown in italics. The polymorphisms relative to the rcrs at the M vertex are: 73G, 263G, 489C, 750G, 1438G, 2706G, 3106del, 4769G, 7028T, 8701G, 8860G, 9540C, 10398G, 10400T, 10873C, 11719A, 12705T, 14766T, 14783C, 15043A, 15301A, 15326G, 16223T.

7 AY Australia DQ Australia AY Australia AF Australia AY Australia AY Australia AY Australia AF PNG highland AF PNG coast AY PNG highland AY PNG highland AY Australian AY Australian AY Australian AY PNG highland AY PNG highland DQ Trobriand Is. DQ Trobriand Is. AY PNG highland AY PNG highland DQ 'Melanesian' P P4 P7 N/R AY AF AY P1 P6 R21 AY AY P G 15613G P 15607G AY AY AY AY P2 P5 AY AY AY AY AY AF R21 P7 P A 11016A 11288T 16274A 16291T 10398A** AF AY AY DQ DQ R G 146C 195C 199C 249G 1709A 1719A 1872C 2397T 4711T 8279C 9067G 9077C 9109G 10610TtvA 11002G 11404G 12234G 12510T 13145A 16168T 16295T 16304C 374G 2283T 6249A 7325G 8387A 10373A 10398A** 10810C 12026G 12235C 14502C 15607G 16148T 16234T 591AtvC 1719A 3010A 5075C 5237A 5276G 6881G 8563G 10088T 12007A 16258CtvA 16263C 16337T 548T 574G 1619T 3987G 5460A 7337A 7340A 8158G 8389G 8749C 10192AtvC 10933G 15244G 15884CtvG 16249C 16312G 16371G P6 271T 4745G 4976G 7010T 8152A 10034C 11176A 11348T 11928G 13545T 14668T 14871C 14923T 14947T 15378C 15930A 16172C 16193T 16245T 16278T 16526A 6719C 16311C 207A 3010A 3306T 6632C 7805A 8562T 8577G 12795A 12973T 13135A 14308C 14385AtvC 16148T 16170CtvA 16256T 16270T 16355T 523insCA 5330AtvC 10790C 12850G 13722G 14470C 14971C 16111AtvC 16169T 16294T 125C 127C 195T 212T(2) 339G 2359T 4017T 8865A 11473G 11807G 12603T 12842C 13105G 15479C 16287T 16294T 212C(1) 6077T P C 16266T 16357C 13269G 16176T 629C 8485A 523insCA 8841T 1719A 12338C 6324A 12346T 7645C 12414C 8764A 15133G 10253C 15514C 11611A 16291T 13927TtvA 16292T 14097T 16335G 16439AtvC 6366A 16235G(1) 480C 4994G 13641C 14374C 15924G 16183CtvA 16223T 3699T 8269A 12346T 12879C 16257T 16270T 16354T 523insCA 9257T 14180C 16235A(2) P3 236C 942G 1834C 2010C 5972T 6253C 6770G 6929G 12121C 13020C 13641C 13708A 13934T 15061G 15454C 127C 128T 567ins3C 3645C 6734A 14338T 15748C 15937TtvA 16145A 522,523CA(2) 1048T 4659A 5420C 6378C 8075A 15077A 15090C 15191C 15889C 16297C 125C 5895delC 13651G 16278T 567ins3C 2056A 4218C 8383C 11339C 11864C 14767C 15523T 16519T(2) (1) P5 7419A 7657C 9509C 12406A 12831T 13368A 15412C 15970C 16192T 16311C P2 709A 4384C 8578T 9293T 10118C 15355A 1438A 3203G 3882A 4122G 8859T 14890G 103A 143A 1375T 16188T 16278T 16357C 16497G 8572A 4892T 16468C 5423G 11914A 12026G AY Malaysia Supplementary Figure 4: P haplogroup with R21 minimal tree and branch-labelled a) The single most parsimonious minimal tree found, from 19 P individuals with the single R21 sequence and L3 outgroup (entire mtdna sequence, no. of parsimony informative chars=64, parsimony score=93). Sequences from this study are shown in bold type. b) A base-labelled reconstructed from the minimal tree with Kivisild et al (2006) P coding-region sequences and partial R12 sequence added (broken lines). Length variations in the poly-c region from nt303-nt315 are not shown. Other gapped sites in the alignment (for example at the short 'CA' microsatellite region from n514-n523) were not included in the analysis, but are shown on the labelled tree. Substitutions are transitions to the base shown, unless marked 'tvn' where N is the base in the rcrs. Synonymous transitions are shown in red type, and sites which change more than once in a lineage in the labelled are followed by the number of the change in brackets. When a substitution results in the same nucleotide as in the rcrs this is shown in italics. Note that nt10398a, labelled with two asterisks in the diagram, is grouping P subclades P1, P2, P3, P5, P6 separately to P4 in this analysis. The n10398a variant is considered to be ancestral to the N macrohaplogroup, reverting to10398g in parallel in several descendant N lineages; including P and R21, resulting in the shown above. It is more probable that the P subclades branch directly from the P vertex, with a reversion to 10398G in P4, as shown in Friedlaender et al 2005 (Mol Biol Evol 22: ). The polymorphisms relative to the rcrs at the N/R vertex in the labelled are: 73G, 263G, 750G, 1438G, 2706G, 3106del, 4769G, 7028T, 8860G, 10398G, 11719A, 14766T, 15326G.

8 DQ Melanesian DQ Melanesian DQ New Britain DQ New Britain DQ New Britain AY PNG highland AY PNG coast DQ Melanesian AY PNG coast AY Bougainville AY PNG highland AY PNG coast AY West New Britain AY West New Britain AY West New Britain AF PNG coast DQ Vanuatu DQ Vanuatu DQ Samoa DQ Cook Islands AY PNG highland AY PNG coast DQ Melanesian DQ Melanesian DQ Melanesian a Q M G 9959C b M29 151T 211G 310C 464G 513A 1048T 4050T 4230T 5263T 7711C 9852G 12127A 12366G 13065T 13452T 16182CtvA 16183CtvA 523insCACA M 13500C 16311C(1) 4025T 7169C 13368A 15939T 10790C 11260C 143A 2768G 4335T 15172A(1) 59C 62TtvG 64T 4023G 6260A 7028C 8531G 8842G 9755A 11914A 12530G 13269G 14209G 16242AtvC Q 3506T 5417A 12248G 15172G(2) 4117C 5843G 8790A 12940A 16129A 16241G(1) 9254G 4639C 61T 62A 453C 5177A 11963A 13641C 13681G 16209C 227TtvA 5351G 16320T 152T 4787G 4917G 6260A 12675T 16390A 195C 228TtvG 5557C 10214T 10283G 11061T 15519C 16066G 16311T(2) 198T 1809C 5821A 6881G 12354insA 12768G 15883A 5460A Q3 Q2 Q1 474C 4679C 8638G 9776T 9941G 12324N 16176T 199C 3394C 3999C 16093C 16209C 16221AtvC 92A(1) 593G 8573A 89C 146C 8964T 14025C 16144C 16148T 16265CtvA 16343G 16488T 1391C 4707T 6221C 7388G 10801A 14404T 13368A 14831R AY AY AY AF DQ DQ C 3398C 4674G 8152A 8910T 9266A 9770C 16187T 16222T 7681T Q3 DQ DQ G(2) 15498A Q1 1407AtvT 3834A 4913CtvA 9101C 11884G 13047G 14798C 16526A 16293G Q2 AY DQ DQ DQ AY AY C 207A 1375T 7993C 10256C 11314G 16241A(2) AY AY AY AY C 861C AY C 8577G 10142T 12477C 14070TtvA 16255A Supplementary Figure 5: Q haplogroup with M29 consensus network and branch-labelled a) The consensus tree of 2 most parsimonious minimal trees found, from 16 Q and 3 M29 individuals with L3 outgroup (entire mtdna sequence, no. of parsimony informative chars=78, parsimony score=92). Sequences from this study are shown in bold type. b) A branch-labelled of the Q and M29 haplogroups reconstructed from the consensus network, with Kivisild et al (2005) coding-region sequences added (broken lines). Length variations in the poly-c region from nt303-nt315 are not shown. Other gapped sites in the alignment (for example at the short 'CA' microsatellite region from n514-n523) were not included in the analysis, but are shown on the labelled tree (note that as the insertion at n12354 in AY causes a frameshift mutation at the 7th codon of ND5 it is likely to a sequencing error). Substitutions are transitions to the base shown, unless marked 'tvn' where N is the base in the rcrs. Synonymous transitions are shown in red type, and sites which change more than once in a lineage in the labelled are followed by the number of the change in brackets. When a substitution results in the same nucleotide as in the rcrs this is shown in italics. The polymorphisms relative to the rcrs at the M vertex are: 73G, 263G, 489C, 750G, 1438G, 2706G, 3106del, 4769G, 7028T, 8701G, 8860G, 9540C, 10398G, 10400T, 10873C, 11719A, 12705T, 14766T, 14783C, 15043A, 15301A, 15326G, 16223T.