A new method called the neighbor-joining method is proposed for reconstructing phylogenetic trees from evolutionary distance data. The principle of this method is to find pairs of operational taxonomic units (OTUs [= neighbors]) that minimize the total branch length at each stage of clustering of OTUs starting with a starlike tree. The branch lengths as well as the topology of a parsimonious tree can quickly be obtained by using this method. Using computer simulation, we studied the efficiency of this method in obtaining the correct unrooted tree in comparison with that of five other tree-making methods: the unweighted pair group method of analysis, Farris's method, Sattath and Tversky's method, Li's method, and Tateno et al.'s modified Farris method. The new, neighbor-joining method and Sattath and Tversky's method are shown to be generally better than the other methods.

Human-chimpanzee comparative genome research is essential for narrowing down genetic changes involved in the acquisition of unique human features, such as highly developed cognitive functions, bipedalism or the use of complex language. Here, we report the high-quality DNA sequence of 33.3 megabases of chimpanzee chromosome 22. By comparing the whole sequence with the human counterpart, chromosome 21, we found that 1.44% of the chromosome consists of single-base substitutions in addition to nearly 68,000 insertions or deletions. These differences are sufficient to generate changes in most of the proteins. Indeed, 83% of the 231 coding sequences, including functionally important genes, show differences at the amino acid sequence level. Furthermore, we demonstrate different expansion of particular subfamilies of retrotransposons between the lineages, suggesting different impacts of retrotranspositions on human and chimpanzee evolution. The genomic changes after speciation and their biological consequences seem more complex than originally hypothesized.

The efficiency of obtaining the correct tree by the maximum likelihood method (Felsenstein 1981) for inferring trees from DNA sequence data was compared with trees obtained by distance methods. It was shown that the maximum likelihood method is superior to distance methods in the efficiency particularly when the evolutionary rate differs among lineages.

Pseudomonas syringae are differentiated into approximately 50 pathovars with different plant pathogenicities and host specificities. To understand its pathogenicity differentiation and the evolutionary mechanisms of pathogenicity-related genes, phylogenetic analyses were conducted using 56 strains belonging to 19 pathovars. gyrB and rpoD were adopted as the index genes to determine the course of bacterial genome evolution, and hrpL and hrpS were selected as the representatives of the pathogenicity-related genes located on the genome (chromosome). Based on these data, NJ, MP, and ML phylogenetic trees were constructed, and thus 3 trees for each gene and 12 gene trees in total were obtained, all of which showed three distinct monophyletic groups: Groups 1, 2 and 3. The observation that the same set of OTUs constitute each group in all four genes suggests that these genes had not experienced any intergroup horizontal gene transfer within P. syringae but have been stable on and evolved along with the P. syringae genome. These four index genes were then compared with another pathogenicity-related gene, argK (the phaseolotoxin-resistant ornithine carbamoyltransferase gene, which exists within the argK-tox gene cluster). All 13 strains of pv. phaseolicola and pv. actinidiae used had been confirmed to produce phaseolotoxin and to have argK, whose sequences were completely identical, without a single synonymous substitution among the strains used (Sawada et al. 1997a). On the other hand, argK were not present on the genomes of the other 43 strains used other than pv. actinidiae and pv. phaseolicola. Thus, the productivity of phaseolotoxin and the possession of the argK gene were shown at only two points on the phylogenetic tree: Group 1 (pv. actinidiae) and Group 3 (pv. phaseolicola). A t test between these two pathovars for the synonymous distances of argK and the tandemly combined sequence of the four index genes showed a high significance, suggesting that the argK gene (or argK-tox gene cluster) experienced horizontal gene transfer and expanded its distribution over two pathovars after the pathovars had separated, thus showing a base substitution pattern extremely different from that of the noncluster region of the genome.

Statistical methods for computing the standard errors of the branching points of an evolutionary tree are developed. These methods are for the unweighted pair-group method-determined (UPGMA) trees reconstructed from molecular data such as amino acid sequences, nucleotide sequences, restriction-sites data, and electrophoretic distances. They were applied to data for the human, chimpanzee, gorilla, orangutan, and gibbon species. Among the four different sets of data used, DNA sequences for an 895-nucleotide segment of mitochondrial DNA (Brown et al. 1982) gave the most reliable tree, whereas electrophoretic data (Bruce and Ayala 1979) gave the least reliable one. The DNA sequence data suggested that the chimpanzee is the closest and that the gorilla is the next closest to the human species. The orangutan and gibbon are more distantly related to man than is the gorilla. This topology of the tree is in agreement with that for the tree obtained from chromosomal studies and DNA-hybridization experiments. However, the difference between the branching point for the human and the chimpanzee species and that for the gorilla species and the human-chimpanzee group is not statistically significant. In addition to this analysis, various factors that affect the accuracy of an estimated tree are discussed.

Insertions and deletions are responsible for gaps in aligned nucleotide sequences, but they have been usually ignored when the number of nucleotide substitutions was estimated. We compared six sets of nuclear and mitochondrial noncoding DNA sequences of primates and obtained the estimates of the evolutionary rate of insertion and deletion. The maximum-parsimony principle was applied to locate insertions and deletions on a given phylogenetic tree. Deletions were about twice as frequent as insertions for nuclear DNA, and single-nucleotide insertions and deletions were the most frequent in all events. The rate of insertion and deletion was found to be rather constant among branches of the phylogenetic tree, and the rate (approximately 2.0/kb/Myr) for mitochondrial DNA was found to be much higher than that (approximately 0.2/kb/Myr) for nuclear DNA. The rates of nucleotide substitution were about 10 times higher than the rate of insertion and deletion for both nuclear and mitochondrial DNA.

The DNA Data Bank of Japan (DDBJ, http://www.ddbj.nig.ac.jp) has made an effort to collect as much data as possible mainly from Japanese researchers. The increase rates of the data we collected, annotated and released to the public in the past year are 43% for the number of entries and 52% for the number of bases. The increase rates are accelerated even after the human genome was sequenced, because sequencing technology has been remarkably advanced and simplified, and research in life science has been shifted from the gene scale to the genome scale. In addition, we have developed the Genome Information Broker (GIB, http://gib.genes.nig.ac.jp) that now includes more than 50 complete microbial genome and Arabidopsis genome data. We have also developed a database of the human genome, the Human Genomics Studio (HGS, http://studio.nig.ac.jp). HGS provides one with a set of sequences being as continuous as possible in any one of the 24 chromosomes. Both GIB and HGS have been updated incorporating newly available data and retrieval tools.

There are three common alleles (A, B, and O) at the human ABO blood group locus. We compared nucleotide sequences of these alleles, and relatively large numbers of nucleotide differences were found among them. These differences correspond to the divergence time of at least a few million years, which is unusually large for a human allelic divergence under neutral evolution. We constructed phylogenetic networks of human and nonhuman primate ABO alleles, and at least three independent appearances of B alleles from the ancestral A form were observed. These results suggest that some kind of balancing selection may have been operating at the ABO locus. We also constructed phylogenetic trees of ABO and their evolutionarily related alpha-1,3-galactosyltransferase genes, and the divergence time between these two families was estimated to be roughly 400 MYA.

A mathematical theory for computing the probabilities of various nucleotide configurations among related species is developed, and the probability of obtaining the correct tree (topology) from nucleotide sequence data is evaluated using models of evolutionary trees that are close to the tree of mitochondrial DNAs from human, chimpanzee, gorilla, orangutan, and gibbon. Special attention is given to the number of nucleotides required to resolve the branching order among the three most closely related organisms (human, chimpanzee, and gorilla). If the extent of DNA divergence is close to that obtained by Brown et al. for mitochondrial DNA and if sequence data are available only for the three most closely related organisms, the number of nucleotides (m*) required to obtain the correct tree with a probability of 95% is about 4700. If sequence data for two outgroup species (orangutan and gibbon) are available, m* becomes about 2600-2700 when the transformed distance, distance-Wagner, maximum parsimony, or compatibility method is used. In the unweighted pair-group method, m* is not affected by the availability of data from outgroup species. When these five different tree-making methods, as well as Fitch and Margoliash's method, are applied to the mitochondrial DNA data (1834 bp) obtained by Brown et al. and by Hixson and Brown, they all give the same phylogenetic tree, in which human and chimpanzee are most closely related. However, the trees considered here are "gene trees," and to obtain the correct "species tree," sequence data for several independent loci must be used.

The nucleotide sequences of four genes of the influenza A virus (nonstructural protein, matrix protein, and a few subtypes of hemagglutinin and neuraminidase) are compiled for a large number of strains isolated from various locations and years, and the evolutionary relationship of the sequences is investigated. It is shown that all of these genes or subtypes are highly polymorphic and that the polymorphic sequences (alleles) are subject to rapid turnover in the population, their average age being much less than that of higher organisms. Phylogenetic analysis suggests that most polymorphic sequences within a subtype or a gene appeared during the last 80 years and that the divergence among the subtypes of hemagglutinin genes might have occurred during the last 300 years. The high degree of polymorphism in this RNA virus is caused by an extremely high rate of mutation, estimated to be 0.01/nucleotide site/year. Despite the high rate of mutation, most influenza virus genes are apparently subject to purifying selection, and the rate of nucleotide substitution is substantially lower than the mutation rate. There is considerable variation in the substitution rate among different genes, and the rate seems to be lower in nonhuman viral strains than in human strains. The difference might be responsible for the so-called freezing effect in some viral strains.

The maximum likelihood (ML) method for constructing phylogenetic trees (both rooted and unrooted trees) from DNA sequence data was studied. Although there is some theoretical problem in the comparison of ML values conditional for each topology, it is possible to make a heuristic argument to justify the method. Based on this argument, a new algorithm for estimating the ML tree is presented. It is shown that under the assumption of a constant rate of evolution, the ML method and UPGMA always give the same rooted tree for the case of three operational taxonomic units (OTUs). This also seems to hold approximately for the case with four OTUs. When we consider unrooted trees with the assumption of a varying rate of nucleotide substitution, the efficiency of the ML method in obtaining the correct tree is similar to those of the maximum parsimony method and distance methods. The ML method was applied to Brown et al.'s data, and the tree topology obtained was the same as that found by the maximum parsimony method, but it was different from those obtained by distance methods.

Based on the morphological characteristics of the skull and teeth, Hanihara ([1991] Japan Review 2:1-33) proposed the "dual structure model" for the formation of modern Japanese populations. We examine this model by dividing it into two independent hypotheses: 1) the Upper Paleolithic population of Japan that gave rise to the Neolithic Jomon people was of southeast Asian origin, and 2) modern Ainu and Ryukyuan (Okinawa) populations are direct descendants of the Jomon people, while Hondo (Main Island)-Japanese are mainly derived from the migrants from the northeast Asian continent after the Aeneolithic Yayoi period. Our aim is to examine the extent to which the model is supported by genetic evidence from modern populations, particularly from Japan and other Asian areas. Based on genetic distance analyses using data from up to 25 "classic" genetic markers, we find first that the three Japanese populations including Ainu and Ryukyuan clearly belong to a northeast Asian cluster group. This negates the first hypothesis of the model. Then, we find that Ainu and Ryukyuans share a group contrasting with Hondo-Japanese and Korean, supporting the second hypothesis of the model. Based on these results, we propose a modified version of the dual structure model which may explain the genetic, morphological, and archaeological evidence concerning the formation of modern Japanese populations.

Mitochondrial DNA (mtDNA) polymorphisms were detected using 13 restriction enzymes on the total DNA obtained from blood samples of five Asian populations: Japanese and Ainu of northern Japan, Korean, Negrito (Aeta) of the Philippines, and Vedda of Sri Lanka. Of a total of 28 restriction-enzyme morphs detected, eight had not been reported previously. By combining the morphs, we were able to classify mtDNAs of 243 individuals into 20 mtDNA types. Phylogenetic analyses using maximum parsimony and genetic distance methods both showed that the Japanese, Ainu, and Korean populations were closely related to each other. Aeta was found to show a relatively close relationship to these three populations, confirming the conclusion from previous studies of blood markers. In contrast, Vedda was quite different from the other four populations.

We determined four nucleotide sequences of the hominoid immunoglobulin alpha (C alpha) genes (chimpanzee C alpha 2, gorilla C alpha 2, and gibbon C alpha 1 and C alpha 2 genes), which made possible the examination of gene conversions in all hominoid C alpha genes. The following three methods were used to detect gene conversions: 1) phenetic tree construction; 2) detection of a DNA segment with extremely low variability between duplicated C alpha genes; and 3) a site by site search of shared nucleotide changes between duplicated C alpha genes. Results obtained from method 1 indicated a concerted evolution of the duplicated C alpha genes in the human, chimpanzee, gorilla, and gibbon lineages, while results obtained from method 2 suggested gene conversions in the human, gorilla, and gibbon C alpha genes. With method 3 we identified clusters of shared nucleotide changes between duplicated C alpha genes in human, chimpanzee, gorilla, and gibbon lineages, and in their hypothetical ancestors. In the present study converted regions were identified over the entire C alpha gene region excluding a few sites in the coding region which have escaped from gene conversion. This indicates that gene conversion is a general phenomenon in evolution, that can be clearly observed in non-functional regions.

Type C retroviruses, designated simian T-cell lymphotropic virus type I (STLV-I), have been isolated from several genera of Old World monkeys and apes, but not from New World monkeys and prosimians. To determine the genomic diversity and molecular evolution of STLV-I and to clarify their genetic relationship to human T-cell lymphotropic virus type I (HTLV-I), we enzymatically amplified, then directly sequenced selected regions of the gag, pol, env, and pX genes of STLV-I strains from Asia and Africa. STLV-I strains Si-2, Matsu, and JM86 from Japanese macaques, which exhibited sequence similarities ranging from 98.5 to 99.8% among themselves, diverged by 12.9 to 13.3% from STLV-I strain MM39-83 from a naturally infected rhesus macaque, by 9.7 to 11.2% from STLV-I strains from Africa, and by 8.8 to 11.2% from HTLV-I strains originating in Japan, India, Africa, the Caribbean, the Americas, Polynesia, and Melanesia. By contrast, the interspecies nucleotide sequence similarity among African STLV-I strains from green monkey, yellow baboon, sooty mangabey, and common chimpanzee was remarkably high, ranging from 96.9 to 97.4%, and these STLV-I strains diverged by only 2.2 to 2.8% from HTLV-I strain EL from equatorial Zaire. Phylogenetic trees constructed by using the neighbor-joining and maximum parsimony methods indicated that the Asian STLV-I strains diverged from the common ancestral virus prior to African STLV-I and cosmopolitan and Melanesian HTLV-I strains. Thus, our data are consistent with an archaic presence of STLV-I in Asia, probably predating macaque speciation, with subsequent independent virus evolution in Asia and Africa.

We have cloned murine genomic and complementary DNA that is equivalent to the human ABO gene. The murine gene consists of at least six coding exons and spans at least 11 kilobase pairs. Exon-intron boundaries are similar to those of the human gene. Unlike human A and B genes that encode two distinct glycosyltransferases with different donor nucleotide-sugar specificities, the murine gene is a cis-AB gene that encodes an enzyme with both A and B transferase activities, and this cis-AB gene prevails in the mouse population. Cloning of the murine AB gene may be helpful in establishing a mouse model system to assess the functionality of the ABO genes in the future.

Muscle tissues can be divided into six classes; smooth, fast skeletal, slow skeletal and cardiac muscle tissues for vertebrates, and striated and smooth muscle tissues for invertebrates. We reconstructed phylogenetic trees of six protein genes that are expressed in muscle tissues and, using a newly developed program, inferred the phylogeny of muscle tissues by superimposition of five of those gene trees. The proteins used are troponin C, myosin essential light chain, myosin regulatory light chain, myosin heavy chain, actin, and muscle regulatory factor (MRF) families. Our results suggest that the emergence of skeletal-cardiac muscle type tissues preceded the vertebrate/arthropod divergence (ca. 700 MYA), while vertebrate smooth muscle seemed to evolve independent of other muscles. In addition, skeletal muscle is not monophyletic, but cardiac and slow skeletal muscles make a cluster. Furthermore, arthropod striated muscle, urochordate smooth muscle, and vertebrate muscles except for smooth muscle share a common ancestor. On the other hand, arthropod nonmuscle and vertebrate smooth muscle and nonmuscle share a common ancestor.

The amino acid sequence of Fuc-TIX is very highly conserved between mouse and human. The number of non-synonymous nucleotide substitutions of the Fuc-TIX gene between human and mouse was strikingly low, and almost equivalent to that of the alpha-actin gene. This indicates that Fuc-TIX is under a strong selective pressure of preservation during evolution. The human Fuc-TIX (hFuc-TIX) showed a unique characteristics, i.e. hFuc-TIX was not activated by Mn2+ and Co2+, whereas hFuc-TIV and hFuc-TVI were activated by the cations. The hFuc-TIX transcripts were abundantly expressed in brain and stomach, and interestingly were detected in spleen and peripheral blood leukocytes.

The genetic relationships of seven Japanese and four mainland-Asian horse populations, as well as two European horse populations, were estimated using data for 20 microsatellite loci. Mongolian horses showed the highest average heterozygosities (0.75-0.77) in all populations. Phylogenetic analysis showed the existence of three distinct clusters supported by high bootstrap values: the European cluster (Anglo-Arab and thoroughbreds), the Hokkaido-Kiso cluster, and the Mongolian cluster. The relationships of these clusters were consistent with their geographical distributions. Basing our assumptions on the phylogenetic tree and the genetic variation of horse populations, we suggest that Japanese horses originated from Mongolian horses migrating through the Korean Peninsula. The genetic relationship of Japanese horses corresponded to their geographical distribution. Microsatellite polymorphism data were shown to be useful for estimating the genetic relationships between Japanese horses and Asian horses.

In the past year, we at DDBJ (DNA Data Bank of Japan; http://www.ddbj.nig.ac.jp) collected and released 1,066,084 entries or 718,072,425 bases including the whole chromosome 22 of chimpanzee, the whole-genome shotgun sequences of silkworm and various others. On the other hand, we hosted workshops for human full-length cDNA annotation and participated in jamborees of mouse full-length cDNA annotation. The annotated data are made public at DDBJ. We are also in collaboration with a RIKEN team to accept and release the CAGE (Cap Analysis Gene Expression) data under a new category, MGA (Mass Sequences for Genome Annotation). The data will be useful for studying gene expression control in many aspects.

The class III POU transcription factor genes play an important role in the nervous system. Comparison of their entire amino acid sequences disclosed a remarkable feature of particular mammalian class III POU genes. Alanine, glycine, and proline repeats were present in the mammalian Brain-1 gene, whereas most of these repeats were absent in the nonmammalian homologue. The mammalian Brain-2 gene had alanine, glycine, proline, and glutamine repeats, which were missing in the nonmammalian homologue. The mammalian Scip gene had alanine, glycine, proline, and histidine repeats, but the nonmammalian homologue completely lacked these repeats. In contrast, the mammalian Brain-4 gene had no amino acid repeats like its nonmammalian homologue. The mammalian genes containing the characteristic amino acid repeats had another feature, higher GC content. We found a positive correlation between the GC content and the amino acid repeat ratio. Those amino acids were encoded by triplet codons with relatively high GC content. These results suggest that the GC pressure has facilitated generation of the homopolymeric amino acid repeats.

We extracted DNA from the human remains excavated from the Yixi site ( approximately 2,000 years before the present) in the Shandong peninsula of China and, through PCR amplification, determined nucleotide sequences of their mitochondrial D-loop regions. Nucleotide diversity of the ancient Yixi people was similar to those of modern populations. Modern humans in Asia and the circum-Pacific region are divided into six radiation groups, on the basis of the phylogenetic network constructed by means of 414 mtDNA types from 1, 298 individuals. We compared the ancient Yixi people with the modern Asian and the circum-Pacific populations, using two indices: frequency distribution of the radiation groups and genetic distances among populations. Both revealed that the closest genetic relatedness is between the ancient Yixi people and the modern Taiwan Han Chinese. The Yixi people show closer genetic affinity with Mongolians, mainland Japanese, and Koreans than with Ainu and Ryukyu Japanese and less genetic resemblance with Jomon people and Yayoi people, their predecessors and contemporaries, respectively, in ancient Japan.

The noncoding region between tRNAPro and the large conserved sequence block is the most variable region in the mammalian mitochondrial DNA D-loop region. This variable region (ca. 270 bp) of four species of Equus, including Mongolian and Japanese native domestic horses as well as Przewalskii's (or Mongolian) wild horse, were sequenced. These data were compared with our recently published Thoroughbred horse mitochondrial DNA sequences. The evolutionary rate of this region among the four species of Equus was estimated to be 2-4 x 10(-8) per site per year. Phylogenetic trees of Equus species demonstrate that Przewalskii's wild horse is within the genetic variation among the domestic horse. This suggests that the chromosome number change (probably increase) of the Przewalskii's wild horse occurred rather recently.

We sequenced exon 2 of the major histocompatibility complex (MHC) class II DRB3 gene from 471 individuals in four different Japanese populations of cattle (201 Japanese Black, 101 Holstein, 100 Japanese Shorthorn, and 69 Jersey cattle) using a new method for sequence-based typing (SBT). We identified the 34 previously reported alleles and four novel alleles. These alleles were 80.0-100.0% identical at the nucleotide level and 77.9-100.0% identical at the amino acid level to the bovine MHC (BoLA)-DRB3 cDNA clone NR1. Among the 38 alleles, eight alleles were found in only one breed in this study. However, these alleles did not form specific clusters on a phylogenetic tree of 236-base pairs (bp) nucleotide sequences. Furthermore, these breeds exhibited similar variations with respect to average frequencies of nucleotides and amino acids, as well as synonymous and non-synonymous substitutions, in all pairwise comparisons of the alleles found in this study. By contrast, analysis of the frequencies of the various BoLA-DRB3 alleles in each breed indicated that DRB3*1101 was the most frequent allele in Holstein cattle (16.8%), DRB3*4501 was the most frequent allele in Jersey cattle (18.1%), DRB3*1201 was the most frequent allele in Japanese Shorthorn cattle (16.0%) and DRB3*1001 was the most frequent allele in Japanese Black cattle (17.4%), indicating that the frequencies of alleles were differed in each breed. In addition, a population tree based on the frequency of BoLA-DRB3 alleles in each breed suggested that Holstein and Japanese Black cattle were the most closely related, and that Jersey cattle were more different from both these breeds than Japanese Shorthorns.