Ludwig Som
Ludwig Som
org/cgi/content/full/324/5926/485/DC1
                                                                                            1
Supplementary online material
Samples
Samples were obtained from Siberia, Middle and Eastern Europe, China and the Iberian
Peninsula from different time points and cultural horizons (Table S1).
Ancient DNA Extraction and Amplification
DNA was extracted from 152 samples using between 250 mg and 400 mg bone material.
External surfaces of bones were removed by abrasion to minimize environmental
contaminations. Each sample was ground to powder with a freezer mill and incubated in
0.45 M EDTA (pH 8.0) and 0.25 mg/ml proteinase K overnight at room temperature
under rotation. After centrifugation for 5 min at 4,000 rpm in a Universal 320 centrifuge
(Hettich), DNA was purified from the supernatant using a silica based method as
previously described (S1).
SNP amplifications were performed using multiplex PCR (S2, S3). PCR products for
SNPs varied in length between 52 bp and 78 bp (including primers; see Table S3). Four
microliters of extract were used for each multiplex PCR. The initial multiplex PCR was
performed in a 20 µl reaction volume containing 1x AmpliTaq Gold PCR buffer II (ABI),
4 mM MgCl2, 1 mg/ml Bovine Serum Albumin (BSA), 250 µM of dATP, dCTP and
dGTP, 500µM of dUTP, 150 nM of each primer and 2 U of AmpliTaq Gold (ABI). For
each sample, one PCR was performed by adding 1 U of heat-labile Uracil-DNA
Glycosylase (USB) and an initial incubation step of 15 min at 37 °C to control for carry-
over contamination. PCR products were diluted 1/30 and 5 µl (total reaction volume 20
µl) were used for the next step. Singleplex PCRs contained 1x AmpliTaq Gold PCR
Buffer II, 4 mM MgCl2, 1 mg/ml bovine serum albumin (BSA), 250 µM of dATP, dCTP
and dGTP, 500µM of dUTP, 1.5 µM of each primer and 0.5 U of AmpliTaq Gold DNA
polymerase. Both times, PCR was run under the following conditions: denaturation and
Taq activation at 94 °C for 9 min; 30 up to 35 cycles at 95 °C 20 sec; annealing
temperature depending on the primer pair (see Table S3) 30 sec; 72 °C 30 sec and final
extension 4 min 72 °C. Negative extraction controls and negative PCR controls were
used in each PCR. Amplification products were visualized on agarose gels.
Mutation analysis
Modern horses show a high variability in coat coloration. However, no color is confined
to a single breed and the same mutation is responsible for a certain color variant across
breeds, supporting the idea that mutations producing the color variants occurred prior to
breed formation during the domestication process. Unfortunately, it is currently not
possible to describe the coat color phenotype of ancient horses completely as until today
not all genes associated with coat coloration were identified. Thus for some color
phenotypes of interest (e.g. dun), there is as yet no genetic information available. We
used a set of eight SNPs in six genes (Table S2) for detecting basic coat colors (bay,
black and chestnut), two kinds of dilution (silver and cream) and three spotted or painted
colorations (overo, tobiano and sabino). SNPs were chosen according to previous studies
on modern horse populations (S4-S9); primers are listed in Table S3. Biotinylated PCR
products were prepared at the PyroMark Vacuum Prep Workstation according to the
manufacturer’s instructions. Amplicons for each SNP were sequenced using
                                                                                        2
pyrosequencing TM technology on a PSQTM 96MA (Biotage). The SNPs were identified
using PSQTM 96MA and automatically edited by the PSQTM 96MA SNP software. Due
to the large deletion (11bp) of the mutated non–black allele (A) of ASIP, the risk to lose
the mutated allele is relatively high for ancient DNA. For this reason, the number of
replications (n=6) was increased in order to reduce the risk of false homozygote
individuals below 1‰. Additionally, allelic separation of ASIP was verified on page gels.
The results for the color determination are summarized in Tables S4 and S5.
Allelic dropout
The probability P of a false heterozygote individual is calculated after n replicates: P = K
x (K/2)n-1 where K is the observed number of allelic dropouts divided by all heterozygous
individuals (S10). For all genes excepting ASIP we did a minimum of four replications
which reduced the risk of non-detection of a heterozygote individual to an average of 0.3
% (P = 0.0078 for KIT13, 0.0015 for KIT16 and MATP, and 0.00012 for MC1R).
Estimating the allele frequency of missed alleles
We computed the upper bound of the allele frequency of a coat color allele having been
present in the pre-domestication population but not observed in our samples assuming a
binomial sampling. Given n samples, we computed the maximum frequency f5% of a
color allele in the population so that the probability of sampling none of them is above
5% (Table S7).
Estimation of selection coefficient
We estimated the selection coefficient for each locus using Bollback et al.’s method
(S11). This method has been developed to jointly estimate the selection coefficient and
the effective population size from time-series data of allele frequencies. We binned the
Siberian East/Central European and Chinese samples into five time periods with means
about 13,100, 3,700, 2,800, 600 and 200 BC respectively. We assumed a generation time
of five years.. We used default values for all parameters of the program except for the
grid size (Ngrid) of the allelic frequencies that was set to 2000. The likelihood was
computed for selection coefficient ranging from -0.5 to 0.5 and effective population size
ranging from 10,000 to 100,000. All the loci considered are assumed to be independent,
in particular we considered only SILV9 because SILV11 seems to be in complete linkage
disequilibrium with SILV9. See Figure S3 for the change in allelic frequency for each
locus and Table S8 for the estimates of the selection parameter.
Note that, for our dataset, the maximum likelihood values are sensitive to the grid size
but that the confidence intervals seem to be robust to this parameter. Therefore we used a
bigger grid than the one suggested in (S11).
Like the above estimate of maximum allele frequencies of missed alleles, these analyses
are limited by the available sample size. This is certainly also true for our sampling of
color variations for early domesticated horses. Thus, it is possible that we missed
extremely rare variants at a certain point in time. However, detection of additional color
variants at the beginning of domestication would only strengthen our conclusions but not
change the overall picture.
                                                                                          3
Supplementary figures
Figure S1: Map of Eurasia showing the archaeological sites (black dots) from which the
horse samples that were successfully analyzed for SNPs originate. The diagrams show
the proportion of the different coat color phenotypes observed. Numbers show the total
number of samples showing the indicated phenotype. From top to bottom: samples from
the Pleistocene; the Neolithic/Copper Age; the Bronze Age; the Iron Age.
                                                                                    4
5
Fig. S2: Timeline for the first occurrence of the observed coat color phenotypes in
Siberia, Europe and the Iberian Peninsula, respectively. Above the timescale, the timing
of the first depictions of various horse-related implements is shown. Note that the
timeline is interrupted between 12,000 and 5,000 BC.
                                                                                      6
Figure S3: Change in allelic frequency for each locus. The number of chromosomes for
each time point is 22, 20, 20, 36, 38 from left to right (see Table S5). The earliest
Pleistocene samples were excluded. Note that for two genes (ASIP and MC1R), the allele
frequencies change drastically.
                                                                                    7
Supplementary tables
Table S1: Samples analyzed for this study. The samples highlighted in grey are the
samples that gave a complete and reproducible genotype for the coat coloration (Typing).
The extraction and amplification of the samples (Ext/Amp) and the reproduction (Rep)
were performed in two different institutes by Melanie Pruvost (MP) and Michael
Cieslack (MC) at the Humboldt University in Berlin and Sebastian Lippold (SL) and
Melanie Pruvost (MP) at the Max Planck Institute in Leipzig. The Pleistocene samples
are not directly dated, but estimated from context should be around 20,000 years old. The
remaining dates are either calibrated carbon dates or derived from archaeological context.
Siberia
                                                                                                   Geographical
                    Sample    Ext/Amp   Rep.   Det.                  Excavation                                             Date     Typing
                                                                                                     location
                    SP1181A   MP/MP            W(ild)           Maliy Lyakhovsky Isl.              North Siberia       Pleistocene    Yes
                    SP1181B   MP/MP             W              Bol'shoy Lyakhovsky Isl.            North Siberia       Pleistocene    Yes
 PLEISTOCENE
                                                                                                                       8
                              Arz 2‐10    CW/MP      MP       D                Arzan2                    South Siberia (Tuva)      619‐608 BC    Yes
                              Arz 2‐11    CW/MP      MP       D                Arzan2                    South Siberia (Tuva)      619‐608 BC    Yes
                              Arz 2‐12    CW/MP      MP       D                Arzan2                    South Siberia (Tuva)      619‐608 BC    Yes
                              Arz 2‐13    CW/MP      MP       D                Arzan2                    South Siberia (Tuva)      619‐608 BC    Yes
                              Arz 2‐14    CW/MP      MP       D                Arzan2                    South Siberia (Tuva)      619‐608 BC    Yes
                              Bars1A      CW/MP               D              Barsucij Log                South Siberia (Tuva)     370‐150 BC     Yes
                              Bars1B      CW/MP               D              Barsucij Log                South Siberia (Tuva)      400‐200 BC    Yes
                              OKG 001      MP/MP              D        Olon‐Kurin‐Gol 10                  Siberia (Mongolia)       400‐300 BC    Yes
                              OKG 002      MP/MP              D        Olon‐Kurin‐Gol 10                  Siberia (Mongolia)       400‐300 BC    Yes
                              OKG 003      MP/MP     SL       D        Olon‐Kurin‐Gol 10                  Siberia (Mongolia)       400‐300 BC    Yes
                                                                                                                                   9
             ORL2      MP/MP            W           Orlovka               Moldova            4000 BC         No
             ORL3      MP/MP            W           Orlovka               Moldova            4000 BC         No
             ORL4      MP/MP            W           Orlovka               Moldova            4000 BC         No
             CAS1      MP/MP            W         Cascioarele             Romania          3700‐3380 BC      Yes
             MAY1      MP/MP            W           Mayaki                Ukraine          3600‐3100 BC      No
             MAY2      MP/MP            W           Mayaki                Ukraine          3600‐3100 BC      No
             MAY3      MP/MP            W           Mayaki                Ukraine          3640‐3490 BC      Yes
             MAY4      MP/MP            W           Mayaki                Ukraine          3600‐3100 BC      No
             MAY5      MP/MP            W           Mayaki                Ukraine          3250‐3100 BC      Yes
COPPER AGE
China
              Sample   Ext/Amp   Rep.   Det.      Excavation      Geographical location      Date         Typing
              Fen 1     CW/MP               D       Fengtai          China (Qinghai)      905‐800 BC       Yes
IRON
 AGE
                                                                                            10
                          Fen 3       CW/MP              D              Fengtai                China (Qinghai)               1000‐800 BC      Yes
                          Fen 4       CW/MP              D              Fengtai                China (Qinghai)               1000‐800 BC      Yes
Spain
                       Sample     Extraction Rep. Det.                 Excavation                 Geographical location             Date      Typing
                         41        MC/MC      SL   W                    Atxoste                  Iberian Peninsula (Spain)    5500‐4950 BC     No
                         42        MC/MC      SL   W                    Atxoste                  Iberian Peninsula (Spain)    5500‐4950 BC     No
                         43        MC/MC      SL   W                    Atxoste                  Iberian Peninsula (Spain)    5500‐4950 BC     No
                         44        MC/MC      SL   W                    Atxoste                  Iberian Peninsula (Spain)    5500‐4950 BC     Yes
                         45        MC/MC      SL   W                    Atxoste                  Iberian Peninsula (Spain)    5500‐4950 BC     Yes
MESOLITHIC‐NEOLITHIC
                                                                                                                               11
            14   MC/MC   SL   D      Peñalosa‐Jaén ‐Andalucia       Iberian Peninsula (Spain)   2200 ‐ 800 BC   No
            15   MC/MC   SL   D      Peñalosa‐Jaén ‐Andalucia       Iberian Peninsula (Spain)   2200 ‐ 800 BC   No
            18   MC/MC   SL   D    Morra Del Quintanar‐ Albacete    Iberian Peninsula (Spain)   2200 ‐ 800 BC   No
            16   MC/MC   SL   D     La Mota ‐Medina Del Campo       Iberian Peninsula (Spain)   800 BC ‐ 6 AD   No
            4    MC/MC   SL   W?    Soto de Medinilla ‐Valladolid   Iberian Peninsula (Spain)   800 BC ‐ 6 AD   Yes
 IRON AGE
                                                                                                12
Table S2. Genes associated with coat color variation typed in this study. Shown are gene
name, GenBank accession number, position and type of mutation, the associated color of both
the wildtype and the derived state and the reference in which the mutation was first described.
Table S3: Primers used for the coat color SNPs amplification. An annealing temperature of
57 °C was used for all of the primer pairs.
                                                                                                  13
       Pk-SILV-I9 S   GGG GAG TGG GCA GAG             15
ASIP   Pk-ASIP-E2 F   CAA GAA ATC CAA AAA GAT CAG C   22   5’-Biotin   67/78 bp
       Pk-ASIP-E2 R   ATG AGA AGT CCA AGG CCT ACC T   22
       P-ASIP-E2 S    CCT ACC TTG GAA GAT CTC         18
                                                                        14
Table S4. Sample information for the horse fossils successfully typed for the coat color SNPs. The table contains information on geographic
location, excavation, dating, and color phenotype of the samples. Grey shaded dates indicate samples directly dated by radiocarbon dating (see also
Table S6). All radiocarbon ages are given in calibrated years (cal.). The Late Pleistocene samples were estimated at around 20,000 years based on
context.
Siberia
                          Sample    Designation        Location                        Excavation                         Date            Phenotype
                         SP1181A      W(ild)         North Siberia               Maliy Lyakhovsky Isl.              Late Pleistocene         Bay
      PLEISTOCENE
                                                                                                                                             Bay
                          Arz 2-2        D        South Siberia (Tuva)                    Arzan2                      619-608 BC            Black
                          Arz 2-3        D        South Siberia (Tuva)                    Arzan2                      619-608 BC           Chestnut
                          Arz 2-4        D        South Siberia (Tuva)                    Arzan2                      619-608 BC             Bay
                          Arz 2-5        D        South Siberia (Tuva)                    Arzan2                      619-608 BC           Chestnut
                          Arz 2-6        D        South Siberia (Tuva)                    Arzan2                      619-608 BC             Bay
                          Arz 2-7        D        South Siberia (Tuva)                    Arzan2                      619-608 BC            Black
                                                                                                                                                15
                               Arz 2-8    D   South Siberia (Tuva)        Arzan2           619-608 BC         Buckskin
                               Arz 2-9    D   South Siberia (Tuva)        Arzan2           619-608 BC         Buckskin
                                                                                                               Chestnut
                               Arz 2-10   D   South Siberia (Tuva)        Arzan2           619-608 BC
                                                                                                               Tobiano
                               Arz 2-11   D   South Siberia (Tuva)        Arzan2            619-608 BC        Chestnut
                               Arz 2-12   D   South Siberia (Tuva)        Arzan2            619-608 BC        Chestnut
                               Arz 2-13   D   South Siberia (Tuva)        Arzan2            619-608 BC        Chestnut
                               Arz 2-14   D   South Siberia (Tuva)        Arzan2            619-608 BC           Bay
                                Bars1A    D   South Siberia (Tuva)      Barsucij Log      cal. 370-150 BC    Bay Tobiano
                                Bars1B    D   South Siberia (Tuva)      Barsucij Log      cal. 400-200 BC    Bay Tobiano
                               OKG 001    D    Siberia (Mongolia)    Olon-Kurin-Gol 10      400-300 BC        Chestnut
                               OKG 002    D    Siberia (Mongolia)    Olon-Kurin-Gol 10      400-300 BC       Bay Tobiano
                               OKG 003    D    Siberia (Mongolia)    Olon-Kurin-Gol 10      400-300 BC        Buckskin
Eastern Europe
                               PET1       W     South Germany           Petersfels       14000-11000   BC       Bay
     GLACIATION - MESOLITHIC
                                                                                                                   16
                             Pie9    W       Romania            Pietrele            4300 BC          Black
    ENEOLITHIC-COPPER AGE
China
 IRON AGE                   Fen 1    D    China (Qinghai)       Fengtai        cal. 905-800 BC    Bay Tobiano
                            Fen 2    D    China (Qinghai)       Fengtai       cal. 1000-800 BC        Bay
                            Fen 3    D    China (Qinghai)       Fengtai       cal. 1000-800 BC        Bay
                                                                                                        17
                    Fen 4   D    China (Qinghai)               Fengtai                cal. 1000-800 BC      Bay
Iberian Peninsula
                     44     W        Spain                    Atxoste                  5500- 4950 BC        Bay
                     45     W        Spain                    Atxoste                  5500- 4950 BC        Bay
    MESOLITHIC-
     NEOLITHIC
                                                                                                               18
Table S5. Results for SNP typing. In addition to the phenotype, the genotypes for all eight SNPs investigated are shown. Grey shading indicates that
at least one allele differs from the wildtype. We did not detect any variation at the EDNRB locus.
                           TAR 001       D              Bay        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                           TAR 002       D           Chestnut      a/a    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                           TAR 004       D              Bay        A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                           TAR 005       D          Bay Sabino     A/A    ov/ov    KM0/KM0    SB1/sb1    C/C    E/e     z/z     z/z
                           TAR 007       D             Black       a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                           TAR 008       D              Bay        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                           TAR 010       D             Black       a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                           TAR 011       D             Black       a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                           BER 002       D              Bay        A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                            Arz 1-2      D           Chestnut      A/a    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                            Arz 1-3      D          Black Silver   a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     Z/z     Z/z
                            Arz 2-1      D              Bay        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
        IRON AGE
                            Arz 2-2      D             Black       a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                            Arz 2-3      D           Chestnut      A/A    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                            Arz 2-4      D              Bay        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                            Arz 2-5      D           Chestnut      A/a    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                            Arz 2-6      D              Bay        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                            Arz 2-7      D             Black       a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                            Arz 2-8      D           Buckskin      A/a    ov/ov    KM0/KM0     sb1/sb1   C/cr   E/e     z/z     z/z
                                                                                                                                                 19
                                Arz 2-9       D            Buckskin        A/a    ov/ov    KM0/KM0     sb1/sb1   C/cr   E/e     z/z     z/z
                               Arz 2-10       D         Chestnut Tobiano   A/a    ov/ov    KM0/KM1     sb1/sb1   C/C     e/e    z/z     z/z
                               Arz 2-11       D            Chestnut        A/A    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                               Arz 2-12       D            Chestnut        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                               Arz 2-13       D            Chestnut        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                               Arz 2-14       D               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                                Bars1A        D          Bay Tobiano       A/a    ov/ov    KM0/KM1     sb1/sb1   C/C    E/e     z/z     z/z
                                Bars1B        D          Bay Tobiano       A/a    ov/ov    KM0/KM1     sb1/sb1   C/C    E/e     z/z     z/z
                               OKG 001        D            Chestnut        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                               OKG 002        D          Bay Tobiano       A/a    ov/ov    KM0/KM1     sb1/sb1   C/C    E/E     z/z     z/z
                               OKG 003        D            Buckskin        A/a    ov/ov    KM0/KM0     sb1/sb1   C/cr   E/e     z/z     z/z
Eastern Europe
                               Sample     Designation    Phenotype                                   nuclear Genes
                                                                           ASIP   EDNRB     KIT13       KIT16    MATP   MC1R   SILV9   SILV11
                                PET1          W               Bay          A/A     ov/ov   KM0/KM0     sb1/sb1    C/C    E/E     z/z     z/z
     GLACIATION - MESOLITHIC
                                PET2          W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                PET3          W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                PET5          W               Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                PET6          W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                 Kg1          W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                 Kg2          W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                 Kg3          W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                 Kg4          W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                 Kg5          W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                Spa 1         W               Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                Pie9          W              Black         a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                                CAS1          W              Black         a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                VIT2          W              Black         a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                                                                                                                                20
                      VIT4        W             Black         a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
   COPPER AGE
   ENEOLITHIC-
                     MAY3         W              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                     MAY5         W              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                     MAY6         W              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                     MAY7         W              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                     MAY10        W              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                     MOL5         W?            Black         a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                      Gar2        D             Black         a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
                      Gar3        D              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                      Gar4        D              Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                     GRO11        D             Black         a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
        BRONZE AGE
                      Shi 1       D           Chestnut        A/A    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                      Lch 1       D         Chestnut Sabino   A/a    ov/ov    KM0/KM0    SB1/sb1    C/C     e/e    z/z     z/z
                      Lor 1       D              Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                      Mic1        D         Bay Tobiano       A/a    ov/ov    KM0/KM1     sb1/sb1   C/C    E/E     z/z     z/z
                      Mic2        D           Chestnut        A/a    ov/ov    KM0/KM0     sb1/sb1   C/C     e/e    z/z     z/z
                      Mic3        D             Black         a/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                      Mic4        D              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                      Mic5        D          Bay Sabino       A/a    ov/ov    KM0/KM0    SB1/sb1    C/C    E/e     z/z     z/z
China
                     Sample   Designation   Phenotype                                   nuclear Genes
                                                              ASIP   EDNRB     KIT13       KIT16    MATP   MC1R   SILV9   SILV11
                     Fen 1        D         Bay Tobiano       A/A     ov/ov   KM0/KM1     sb1/sb1    C/C    E/E     z/z     z/z
                     Fen 2        D              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/e     z/z     z/z
  IRON AGE
                     Fen 3        D              Bay          A/a    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                     Fen 4        D              Bay          A/A    ov/ov    KM0/KM0     sb1/sb1   C/C    E/E     z/z     z/z
                                                                                                                                   21
Iberian
Peninsula
                      44          W      Bay      A/a   ov/ov   KM0/KM0   sb1/sb1   C/C   E/E   z/z   z/z
                      45          W      Bay      A/a   ov/ov   KM0/KM0   sb1/sb1   C/C   E/E   z/z   z/z
     MESOLITHIC-
      NEOLITHIC
                                                                                                            22
               Table S6. Detailed information on the ages of the samples investigated.
Siberia
                                                                                                                                                  23
                     Arz 2-7    South Siberia (Tuva)        Arzan2           619-608 BC
                     Arz 2-8    South Siberia (Tuva)        Arzan2           619-608 BC
                     Arz 2-9    South Siberia (Tuva)        Arzan2           619-608 BC
                     Arz 2-10   South Siberia (Tuva)        Arzan2           619-608 BC
                     Arz 2-11   South Siberia (Tuva)        Arzan2           619-608 BC
                     Arz 2-12   South Siberia (Tuva)        Arzan2           619-608 BC
                     Arz 2-13   South Siberia (Tuva)        Arzan2           619-608 BC
                     Arz 2-14   South Siberia (Tuva)        Arzan2           619-608 BC
                     Bars1A     South Siberia (Tuva)     Barsucij Log                        2170 ± 30 BP   cal. 370-150 BC   Poz-22611   3.3%N 11.5%C
                     Bars1B     South Siberia (Tuva)     Barsucij Log      cal. 400-200 BC
                    OKG 001      Siberia (Mongolia)    Olon-Kurin-Gol 10     400-300 BC
                    OKG 002      Siberia (Mongolia)    Olon-Kurin-Gol 10     400-300 BC
                    OKG 003      Siberia (Mongolia)    Olon-Kurin-Gol 10     400-300 BC
East Europe
                                                                                                                                                 24
                                Ukraine (Peninsula                                                        cal. 9390-9210
                        Spa 1        Crimea)           Span-Koba                                                BC
                         Pie9       Romania              Pietrele          4300 BC
                                                                                                          cal. 3700-3380                1.0%N 7.6%C
ENEOLITHIC-COPPER AGE
                                                                                                                                              25
China
Iberian Peninsula
                                                                                                                                                                          26
IRON AGE   4    Iberian Peninsula (Spain)   Soto de Medinilla – Valladolid   800 BC - 6 AD
           29   Iberian Peninsula (Spain)      Mucientes - Valladolid                        1275 ± 25 BP   cal. 660-780 AD   Poz-24740   1.2%N 6.9%C
MEDIEVAL
           30   Iberian Peninsula (Spain)      Mucientes - Valladolid                        1230 ± 30 BP   cal. 680-890AD    Poz-24739   3.4%N 15%C
                                                                                                                                                27
Table S7. Upper bound of the frequency f5% a colour allele could have had in the horse
population prior to domestication having a chance to be missed of p ≤ 0.05 (n = number
of alleles).
Population                                  n                     f5%
Siberia                                       10                    0.26
East Europe                                   42                    0.07
Iberia                                        16                    0.15
Siberia + East Europe                         52                    0.06
Siberia + East Europe + Iberia                68                    0.04
Table S8. Maximum likelihood estimates of the selection coefficient for each of the loci.
The results were obtained using Bollback et al’s method (S11), excluding the Pleistocene
samples. To get the 95% confidence interval we used the profile likelihood. For ASIP and
MC1R loci the selection coefficient is significantly different than 0. For the other loci, the
data is compatible with s=0. Therefore, the observed data are difficult to explain by
genetic drift alone under a standard neutral model (for review see S14).
Note that we had no power to estimate the effective population size. Indeed, the
confidence interval for the effective population size spans all the values of the grid we
used to compute the likelihood, that is 10,000 to 100,000. This is not surprising since
Bollback et al. (S11) noted that their method has little power to co-estimate the effective
population size and the selection coefficient.
Supplementary references
S1.   N. Rohland, M. Hofreiter, Nature Protocols 2, 1756 (2007).
S2.   J. Krause et al., Nature 439, 724 (Feb 9, 2006).
S3.   H. Römpler et al., Nat. Protocols 1, 720 (2006).
S4.   S. A. Brooks, E. Bailey, Mamm Genome 16, 893 (Nov, 2005).
S5.   S. A. Brooks, R. B. Terry, E. Bailey, Anim Genet 33, 301 (Aug, 2002).
S6.   E. Brunberg et al., BMC Genet 7, 46 (2006).
S7.   D. Mariat, S. Taourit, G. Guerin, Genet Sel Evol 35, 119 (Jan-Feb, 2003).
S8.   L. Marklund, M. J. Moller, K. Sandberg, L. Andersson, Mamm Genome 7, 895
      (Dec, 1996).
                                                                                           28
S9.    S. Rieder, S. Taourit, D. Mariat, B. Langlois, G. Guerin, Mamm Genome 12, 450
       (Jun, 2001).
S10.   P. Gagneux, C. Boesch, D. S. Woodruff, Mol Ecol 6, 861 (Sep, 1997).
S11.   J. P. Bollback, T. L. York, R. Nielsen, Genetics 179, 497 (May, 2008).
S12.   M. Reissmann, J. Bierwolf, G. Brockmann, Anim Genet 38, 1 (2007).
S13.   E. M. Santschi, A. K. Purdy, S. J.Valberg, P. D. Vrotsos, H. Kaese, J. R.
       Mickelson Mamm Genome 9, 306 (1998).
S14.   N. A. Rosenberg, M. Nordborg Nature Reviews Genetics 3, 380 (May 2002).
29