The distribution of the old samples in the network of current haplotypes represented denotes, at first sight, a great diversity. Among the most divergent haplotypes of the set, identified with the African haplogroup L1b, there are two sequences from the Neolithic era from the Nerja site (Málaga). At one end of the L2 group there are another five samples from three different sites, three of which, from the Chalcolithic site of Tres Montes (Navarra), share the same haplotype. Related to them are the haplotypes of samples 1MA12 and 2H20 from the Syrian sites of Mari (~4500 A.P.) and Tell Halula (10000 A.P.).
The representation of the African heritage is completed by three other samples from Mari (2MA10), Tell Halula (1H15) and Toledo (TO1) that are included, along with other haplotypes, in haplogroup L3. A fork in the clade to which these sequences belong leads to the TR19 sequence, which clusters with two other haplotypes from Asian haplogroup C. Some of the remaining ancient sequences occur in clades made up of haplotypes from the same haplogroup. This happens with samples 2MA2 and 1MA5, present within cluster U4; TR16, TR18 and TR12 within the K; 3MA11, 1TM5, 2NE and AB9 in the J; and DJ1 and H12 in its sister group T. Sample 1PI appears related to sequences from the Saami (haplotype 912), Druze (haplotypes 571-X and 566-U1) and from Austria, France, Finland and Karelia of Russia (haplotype 35), all of them in a polyphyletic clade that has already been mentioned. Samples 1MA1 and AB17 are located within a group that contains sequences from haplogroups H (451), L3a (516) and U3. The haplotypes represented in this group are today present in populations of Northeast and Southeast Europe: Finland, Norway, South Italy and Sicily, as well as in the population of South Africa. The Paleolithic sample from the Nerja 3NE site is grouped with three sequences of haplogroup U3 (745, 88 and 852) and with one to which no haplogroup could be assigned (haplotype 453). Such variants are distributed today in various populations of the Near East such as Turkey, Syria and Israel (Druze) and in nearby populations such as Bulgaria, the North Caucasus and the Nile Valley, although they also appear in Portugal, Sicily and Norway. Samples TH5, TH2, 1H17, 2TM4 and TR14 are located within the heterogeneous clade that includes several subgroups of haplogroup U, close to the group formed by subhaplogroup K. As previously mentioned, the only information provided by the Hypervariable Region I is here insufficient to adequately classify these sequences. Samples TR8 and 1H14 appear within the framework formed by direct derivatives of the consensus haplotype (CRS), which also includes 2PI and AB5. The lack of correspondence between the TR8 and 1H14 haplotypes and the set of current sequences justifies their a priori non-inclusion within haplogroup H.
…
Members of the L2 haplogroup are separated from the central node by transition 16278T, which is consistent with the recent study by SALAS et al. 2002. Within this haplogroup are included 18 haplotypes, of which 3 correspond to ancient samples. Such haplotypes occupy the final end of one of the branches of this haplogroup. The 2H20 haplotype (16223T-16261T-16278T-16294T-16309G) is located at the base of the other two ancient haplotypes and in the present reconstruction comes from node 121 (16223T-16278T-16294T-16309G), present in an individual from the Canary Islands, two from East Africa, four from West Africa and one from the Nile Valley. The 2H20 motif is not unique in the current sequence database but also appears in a West African individual. From this node derive, on the one hand, the haplotype of 1MA12 (16223T-16256T-16261T-16278T-16294T-16309G) and, on the other, the haplotype shared by the individuals of the deposit of Tres Montes 1TM2, TM6 and TM11 (16261T-16278T-94T-16309G). In the latter case, there would have been a reversal of the mutation in position 16223, a position with a high mutation rate according to several studies (MEYER et al. 1999; MALYARCHUK et al. ...
2002). In general, the phylogeny of haplogroup L2 is quite well resolved here, with the sole exception of haplotype 506 (187T-223T-278T-362C), which is located outside the L2 clade closest to that of certain representatives of group L1. Haplogroup L3 is represented by 14 haplotypes. The absence of information on certain non-sequenced positions of the Hypervariable I region and restriction enzymes makes resolution at the subcluster level of this broad haplogroup impossible. This difficulty is even more evident in the attempt to reconstruct the L1 haplogroup and its many variants. In Figure FR71 this haplogroup does not constitute a clade, and the representatives of the subcluster L1b appear, on the one hand, as derivatives of a branch of the haplogroup L3 (haplotype 519), while the lines of L1a, L1c, L1d and L1i come from the central node of the haplogroup L2 (haplotype 237) through intermediates of haplogroups L3 and L2 (haplotypes 200 and 506). The combined effect of a small number of haplotypes within each subgroup together with the absence of information regarding certain diagnostic positions makes it impossible to correctly reconstruct phylogenetics between the different variants. However, the algorithm manages to separate the haplotypes of each subcluster in an effective way, which is more than enough for the main objective of the present doctoral thesis, which aims only to place the old sequences obtained within the framework of current sequences. In this sense, the sequences of the individuals 4NE and 5NE of the Neolithic stratum of the Cave of Nerja are framed within the subhaplogroup L1b. Regarding the rest of the haplotypes of this subcluster, they occupy a basal position since none of them presents the transitions 16187T and 16189C. However, reconstructions of this type made with a large number of current African sequences suggest that the original node of the haplogroup L1 is that containing the mutations 16187T-16189C-16311C. . Taking the latter into account, the only explanation for the sequences of 4NE and 5NE is that they derive from an ancestral node in which the first two transitions have been lost. Such an ancestral haplotype is not present in the database of current sequences, but, as discussed above, its existence in the remote past cannot be ruled out.
…
Many of the African haplotypes are correctly classified into haplogroups. This is the case with subclusters L1a, L1b, L1c and L1d of haplogroup L1, with cluster M1, and with most of the haplotypes of group L2. The subvariants of haplogroup L3, L3a and L3b, appear mixed with each other and related to Asian and Native American haplogroups C and D, the first of which includes the sequence of individual TR19. **Another Asian-American haplogroup, haplogroup X, is not well resolved and its haplotypes appear, on the one hand, related to the African haplotypes of the L2 group and, on the other hand, within the polyphyletic clade that includes the old sequence 1PI**.
As other authors have previously observed (RICHARDS et al. 1996), the algorithm used in the phylogenetic reconstruction of Neighbor-Joining separates efficiently sample haplotypes in the European haplogroups J, T, K and US. On the other hand, it tends to group variants of different origin in the same group, as happens here with the group that includes the IPI sequence, and likewise tends to haplotypes with a common origin. As mentioned, the main cause is the presence of homoplastic substitutions that have arisen in parallel along two lines independent. This phenomenon is especially frequent in markers with a of high mutation, as occurs with microsatellites and with the control region of the mitochondrial DNA. In the case of mtDNA, the phenomena of homoplasy and reversion