Thecompletegenomesequenceofa Neanderthal from the Altai Mountains

Transcription

1 Thecompletegenomesequenceofa Neanderthal from the Altai Mountains Kay Prüfer, Fernando Racimo, Nick Patterson, Flora Jay, Sriram Sankararaman,, Susanna Sayer, Anja Heinze, Gabriel Renaud, Peter H. Sudmant, Cesare de Filippo, Heng Li, Sapan Mallick,, Michael Dannemann, Qiaomei Fu,, Martin Kircher,, Martin Kuhlilm, Michael Lachmann, Matthias Meyer, Matthias Ongyerth, Michael Siebauer, Christoph Theunert, Arti Tandon,, Priya Moorjani, Joseph Pickrell, James C. Mullikin 7, Samuel H. Vohr 8, Richard E. Green 8, Ines Hellmann 9 {, Philip L. F. Johnson,Hélène Blanche, Hoard Cann, Jacob O. Kitzman, Jay Shendure, Evan E. Eichler,, Ed S. Lein, Trygve E. Bakken, Liubov V. Golovanova, Vladimir B. Doronichev, Michael V. Shunkov, Anatoli P. Derevianko, Bence Viola, Montgomery Slatkin, David Reich,,7, Janet Kelso & Svante Pääbo

2 a b mm Vindija Mezmaiskaya Denisova 7 9 Figure Toe phalan and location of Neanderthal samples for hich genome-ide data are available. a, The toe phalan found in the east gallery of Denisova Cave in. Dorsal vie (left image), left vie (right image). Total length of the bone is mm. b, Map of Eurasia shoing the location of Vindija Cave, Mezmaiskaya Cave and Denisova Cave, here Neanderthal samples used here ere found.

3 Chrom 8 Tracts>. cm Papuan Denisova Altai Position (Mb) Chrom Tracts>. cm Papuan Denisova Altai Position (Mb) Figure S. HBD tracts identified in chromosomes 8 and for Papuan (top line, green), Denisova (middle line, black), and Altai (bottom line, pink).

4 Denisova Altai Neanderthal Heterozygosity in tract e+ e- 8e- HBD tract length.cm -. cm.cm - cm > cm 8 Heterozygosity in tract e+ e- e- e- HBD tract length.cm -. cm.cm - cm > cm Tract length (cm) Tract length (cm) Figure S.. Heterozygosity in HBD tracts detected by the scan for Denisova and Altai Neanderthal as a function of the tract length.

5

6 Conclusion. The inbreeding coefficient of the Altai individual is likely to be /8, hich implies that her parents ere double first cousins, grandfather and granddaughter, grandmother and grandson, half siblings, uncle and niece, or aunt and nephe, but that e cannot distinguish among these possibilities using runs of homozygosity on the autosomes. This number provides an upper bound for the inbreeding coefficient as a smaller false positive rate or unobserved heterozygous sites (due to missing data) might decrease the total length of homozygous tracts.

7 Figure S.9. Non-ehaustive illustration of pedigrees that can be ecluded (top, A-D) or not ecluded (bottom, E-H), using X chromosome information. Gray denotes the absence of X sequence coming from the recent common ancestor(s). Other colors denote the potential presence of X sequence coming from the common ancestor(s). Dark blue indicates that both parents might carry X chunks inherited from the same recent common ancestor, thus the individual might be inbred for X. The pedigrees depict cases of the folloing scenarios: offspring of half-siblings (A,E), grandfather-granddaughter (B, F), aunt-nephe (C,G), grandmother-grandson (D), double-first-cousins (H)

8 Background coverage for tracts in [.,] cm Background coverage for Neanderthal (ie coverage not eplained by the recent inbreeding scenarios) Background coverage for Denisova (assuming no recent inbreeding) Background coverage for Papuan (assuming no recent inbreeding) double st cousins grandfather granddaughter* half siblings uncle and niece* Offspring of Figure S. Background coverage for tracts beteen. and cm.

9 Conclusion: The observed background coverage of HBD tracts could be eplained by the presence of background inbreeding in the population. Alternatively, a demographic scenario of random mating ith successive bottlenecks starting after the split from modern humans that induce a very small population size at time of sampling (~ individuals) also provides a reasonable fit to the data. Note that hen a population is very small for a long time the chance of mating beteen distant cousins is not negligible even in case of random mating.

10 Patterns of ancient selection in modern humans around candidate sites Fernando Racimo,,, Martin Kuhlilm, Montgomery Slatkin, Department of Integrative Biology, University of California, Berkeley, CA, USA Ma Planck Institute for Evolutionary Anthropology, Leipzig, Germany. Corresponding author: Associate Editor: X

11 FIG.. Tree representing msms runs to simulate a change in a site that is homozygous ancestral in an archaic human (Pop. B) and rises to fiation in modern humans (Pop. A). t AB =modern-archaic split time. t S =derived allele fiation time.

12 Test statistics for sets of adjacent segregating sites H E, = p i ( p i ) (i. e. not π). i= H M, the frequency of the most common haplotype. H S, evenness of haplotype frequency distribution H I, the inconsistency of the majority haplotype ith the outgroup genotype.

13 Figure S. The mean values of the H E, H M, H S and H I statistics from simulations run under the same parameters ere calculated along indos of kb (=. cm) in a Mb region and divided by their mean value along the entire region.

14 FIG.. Poer to reject neutrality for different statistics under to different selection coefficients and a range of times since fiation, estimated by calculating the proportion of simulations (out of ) that have a value more etreme than 9% of neutral simulations.

15 FIG. 7. Sets of simulations ere run through the ABC pipeline to obtain Bayes factors in favor of selection (versus neutrality) under different knon parameters. The lines sho the proportion of the simulations that have a Bayes factor larger than the specified cutoffs. BF = Bayes factor, s=selection coefficient, t=time since derived allele fiation, in generations.

16 Numbers of sites fied derived in humans (p>.99) and fied ancestral in the high-coverage Neanderthal and Denisovan genomes. Nonsynonymous: 9 Synonymous: Splice: ' UTR: ' UTR: 9 Regulatory motif:

17 FIG.. Density of estimated posterior modes from ABC analyses under the positive selection model, across different genomic classes. The grey bars represent the prior used for each parameter.

18 Table. Modern-human specific changes that lead to an amino acid replacement, affect a splice site or are located in a UTR, and that: ) have Bayes factors > in favor of selection and ) are a good fit (P >.) to the selection model. Position Bayes factor log(s) t S (generations) P neutral P selection Class Gene chr: UTR SLC9A chr: UTR SLC9A chr: UTR SLC9A chr: Splice USP chr: Splice ZCWPW chr: UTR ZCWPW chr: UTR ABHDB chr7: UTR STXA chr8: UTR TMEM7 chr: UTR SLC8A chr: UTR TMPRSS chr: UTR TMPRSS chr: UTR PRDM chr: UTR PRDM chr: UTR PRDM chr: NonSyn PRDM chr: NonSyn NAALADL chr: UTR FAU chr: UTR FAU chr: UTR MRPL9 chr: NonSyn MRPL9 chr: UTR SYVN chr: UTR TRHDE chr: UTR TRHDE chr: NonSyn TTLL chr7: UTR KAT7 chr9: UTR ZNF chr9: UTR ZNF NOTE. Parameters listed are the posterior modes inferred using ABC. We also list the P-values for the fit to the neutral and selection models.

19 Projection analysis (Melinda Yang) is the derived allele frequency in the reference population. At each segregating site in the reference population, assign a eight to that site in the test genome = if the test genome is homozygous ancestral = if the test genome is heterozygous = if the test genome is homozygous derived () is the projection of the test genome on the reference population.

20 If the test genome is a random sample from the reference population () = Reference Test τ /( N ) () = e τ Chen et al. (7, Genetics)

21 Reference Test Bottleneck in population size

22 Bneck in Test Bneck in Ref Bneck in Anc Population divergence time, τ=, years, N=,, years per generation Bottleneck (.) in test and reference,-, years ago Bottleneck (.) in ancestral population 7,-, years ago. Bneck in Test and Ref

23 Reference Test or

24 kya Nm= kya From Test into Reference From Reference into Test

25 Reference Test f=. bottleneck (.)

26 Bneck. in Ref, No Admi Bneck. in Ref, kya No Bneck, kya Population separation τ=, years Bottleneck (.),-, years ago Admiture from test to reference (f=.), years ago

27 Reference Test Ghost

28 Test and reference diverged, years ago Ghost diverged, years ago Admiture rate f=.,, years ago into test or reference, 7, years ago into ancestor

29 Reference Test Ghost

30 . Test to Ref

31 French Han Papuan Dinka Yoruba San CEU reference population

32 Altai Deni Losh CEU reference population

33 French Han Papuan Yoruba Mbuti San YRI reference population

34 Altai Deni Losh YRI reference population