Background: Lipoprotein(a) [Lp(a)] is composed of a low density-lipoprotein (LDL) particle and a glycoprotein molecule known as apolipoprotein(a) [apo(a)]. Apo(a) exists in several differently-sized isoforms and is responsible for the unique properties of Lp(a). Although Lp(a) has been known for the past 40 years its relationship with coronary heart disease (CHD) has not been characterized in sufficient detail. Whether Lp(a) causes CHD is not clear. Furthermore, the role of apo(a) isoform variation and other sources of Lp(a) heterogeneity (e.g., level of oxidized phospholipids) in Lp(a)-disease association has not been determined. Objectives: To characterize in detail the association of circulating Lp(a) levels with the risk CHD To assess the nature of Lp(a)-CHD association using an integrative genetic study To explore the role of Lp(a) heterogeneity in its association with CHD Data sources: 1. The Emerging Risk Factors Collaboration (ERFC) database (36 studies, 127,000 participants) 2. The European Prospective Investigation of Cancer – Norfolk (EPIC-Norfolk) study (2200CHD cases, 2200 controls) 3. The Pakistani Risk of Myocardial Infarction Study (PROMIS) (1800 MI cases and 1800 controls) 4. Systematic quantitative reviews of published epidemiological studies Results: ERFC data - Analyses of cross-sectional data on up to 127,000 participants (predominantly of European descent) demonstrated that Lp(a) is generally not strongly correlated with known CHD risk factors. Weakly positive correlations were observed with LDL-cholesterol, apolipoprotein B100 and fibrinogen. Levels were over 2-fold higher in Blacks compared to Whites. Analyses of available data on repeat measurements in 6600 participants demonstrated that Lp(a) values have very high long-term within-person consistency (regression dilution ratio ~ 0.9). Outcome data involved 9300 incident CHD events, 1900 ischaemic strokes and 8100 nonvascular deaths. The risk ratio for CHD per 1SD higher Lp(a) concentration, adjusted for age, sex, lipids and other conventional vascular risk factors, was 1.13 (95% CI, 1.09-1.18). The corresponding risk ratios for ischaemic stroke and nonvascular death were 1.10 (1.02 – 1.18) and 1.01 (0.98-1.05), respectively. Data were too limited to assess association in nonwhites. PROMIS data – the adjusted odds ratio for MI in South Asians was comparable to that of Europeans. EPIC-Norfolk genetic data - The odds ratio for CHD per 1-SD higher Lp(a) concentration, after adjustment for cardiovascular risk factors, was 1.37 (1.20-1.56). Tagging SNPs rs10455872 and rs11751605 (minor allele frequency: 8% and 18%, respectively) were associated with 207% (95% CI, 188 - 227%) and 38% (31 - 46%) higher Lp(a) concentrations per copy of minor allele, respectively. These SNPs accounted for 35% and 5% of the variation in circulating Lp(a) levels, respectively, and were associated with an odds ratio for CHD of 1.34 (1.14-1.58) and 1.17 (1.04-1.33), respectively. The observed SNP-CHD associations were consistent with expected odds ratios corresponding to the Lp(a) effect of the SNPs. Systematic reviews – meta-analysis of published data from 40 studies (11,300 cases, 47,000 controls) demonstrated that people with smaller apo(a) isoforms have about a 2-fold higher risk of CHD or ischemic stroke than those with larger isoforms. Meta-analysis of published data from 10 studies (1500 cases, 10,200 controls) showed that people in the top third of baseline distribution of oxidized LDL levels have a 1.8-fold higher risk of CHD than those in bottom third. EPIC-Norfolk biomarker data – Levels of oxidized phospholipids were strongly correlated with Lp(a) concentration (r = 0.7, p-value < 0.0001). One SD higher concentration of oxidized phospholipids was associated with an adjusted odds ratio for CHD of 1.31 (1.15-1.49). The risk ratio was no longer significant after adjustment for Lp(a) concentration (1.08; 95% CI, 0.91-1.29). Conclusion: Lp(a) concentration is specifically, continuously and independently associated with the risk of ischaemic vascular outcomes. Available evidence supports the causal role of the particle in CHD. Lp(a) appears to induce vascular damage through causal mechanisms that involve apo(a) isoforms and oxidized phospholipids. A comprehensive study of markers of Lp(a) heterogeneity should help to understand the full impact of Lp(a) on cardiovascular diseases. In addition, further study is needed in nonwhites to assess the relevance of the factor to vascular disease risk in these populations.

Este trabalho teve como objetivo caracterizar e padronizar métodos para a tipagem molecular de marcadores genéticos, para futuro uso em estudos de genética populacional de Plasmodium vivax na Amazônia brasileira. As amostras sanguíneas utilizadas neste estudo foram colhidas no Brasil, Camboja, Sri Lanka e Estados Unidos. Foram selecionados polimorfismos de única base (SNPs) distribuídos ao longo de cromossomos distintos de P. vivax; para estes polimorfismos, sete ensaios de tipagem de SNPs foram padronizados com sucesso. Com a tipagem molecular, foi possível definir haplótipos que caracterizam cada amostra, assim como identificar infecções geneticamente mistas (co-ocorrência de clones distintos na mesma amostra). No entanto, com este conjunto de marcadores não foi possível agrupar amostras de acordo com sua localização geográfica, por estes marcadores não serem suficientemente informativos do ponto de vista genético. Os sete SNPs avaliados, quando comparados a 13 marcadores de DNA microssatélite, revelaram menor proporção de infecções geneticamente mistas e menor diversidade genética. / This study aimed to characterize and standardize methods for molecular typing of genetic markers to be used in the future in studies of population genetics of Plasmodium vivax population in the Brazilian Amazon. Blood samples used in this study were collected in Brazil, Cambodia, Sri Lanka and United States. Single nucleotide polymorphisms (SNPs), distributed over different P. vivax chromosomes were chosen and seven typing assays were sucessfully standardized. Molecular typing of polymorphisms allowed to define haplotypes that characterize each sample, as well as to identify genetically mixed infections (co-occurrence of different clones in the same sample). However, with this markers set, it was not possible to group samples according to their geographical location, because probably they are not sufficiently genetically informative. Compared to 13 microsatellite markers, these seven SNPs revealed a lower proportion of mixed-clone infections and lower genetic diversity.

Através da análise de microssatélites e SNPs, incluindo tanto marcadores neutros como sujeitos a seleção, neste trabalho se demonstrou que: 1) As populações brasileiras de P. vivax (Pv) são mais diversas que as populações simpátricas de P. falciparum (Pf), porém ambas apresentam desequilíbrio de ligação. 2) Os polimorfismos neutros apresentam alta variabilidade ao longo do tempo em ambas as espécies, em contraste com a estabilidade observada nos marcadores sob seleção. 3) As altas taxas de recorrências por Pv na Amazônia brasileira são causadas, em sua maioria, por parasitas geneticamente não relacionados. 4) A recombinação não meiótica contribui substancialmente à diversidade dos domínios repetitivos de PvCSP. 5) Os SNPs nos genes pvmdr1 e pvcrt-o de isolados de Pv resistentes à cloroquina não se associam ao seu fenótipo in vivo. Finalmente, através da genotipagem de 57 SNPs do cromossomo 8 de Pv em 234 isolados do Brasil e da Ásia observou-se que, em Pv, a diversidade e a recombinação mitótica não se associam aos níveis de endemicidade como acontece em Pf e que Pv apresenta estrutura populacional continental e subcontinental no Brasil. / Through the analysis of microsatellites and SNPs, including both neutral and under selection markers, this work showed that: 1) The Brazilian populations of P. vivax (Pv) are more diverse than sympatric P. falciparum (Pf) populations, but both exhibit linkage disequilibrium. 2) For both species, neutral polymorphisms showed high variability over time, contrasting with the stability of under markers under selection. 3) The high rate of Pv recurrences in the Brazilian Amazon region is mostly due to genetically unrelated parasites. 4) Non-meiotic recombination substantially contributes to PvCSP repetitive domain diversity. 5) SNPs at pvmdr1 and pvcrt-o genes of chloroquine resistant isolates of Pv are not associated with the in vivo phenotype. Finally, 57 SNP genotyping of chromosome 8 of Pv among 234 isolates from Brazil and Asia showed that, in Pv, diversity and mitotic recombination are not associated with levels of endemicity as in Pf and that Pv presents continental population structure and subcontinental structure in Brazil.

Trypanosamas de diversas espécies podem infectar mamíferos de interesse econômico em todo o mundo. Trypanosoma vivax, T. evansi, T. equiperdum, T. congolense, T. b. brucei e T. simiae geram importantes doenças em ungulados na África, Ásia e Américas Central e do Sul; enquanto T. theileri e espécies relacionadas são pouco patogênicas. Compreender a epidemiologia e as interações parasita-vector-hospedeiro requer estudos de estrutura populacional, filogeográficos e de diversidade e relações filogenéticas entre isolados. Sequências de microssatelites e dos genes SSUrRNA, gGAPDH, CatL, ITS, SL, 5S, Cytb e ESAG6 mostraram ampla diversidade biológica e diferenciaram genótipos com associação geográfica e restrição de hospedeiros em T. theileri e espécies relacionadas. T. evansi mostrou importante heterogeneidade de sequências no gene ESAG6, enquanto populações de T. vivax muito divergentes foram observadas em regiões preservadas e com transmissão cíclica quando comparadas com a microheterogeneidade biológica de áreas de transmissão mecânica. / Trypanosames can infect diverse species of mammals of economic interest worldwide. Trypanosoma vivax, T. evansi, T. equiperdum, T. congolense, T. brucei brucei and T. simiae cause important diseases in ungulates in Africa, Asia and Central and South America, while T. theileri and related species are of low pathogenicity. Understanding the epidemiology and host-parasite-vector interactions requires studies of population structure, phylogeography and diversity and phylogenetic relationships among isolates. Microsatellite loci and sequences from genes SSUrRNA, gGAPDH, CatL, ITS, SL, 5S, Cytb and ESAG6 revealed high biological diversity and allowed differentiation of genotypes with geographic structure and host-restriction event in T. theileri and related species. T. evansi showed significant heterogeneity in ESAG6 sequences, while populations of T. vivax widely divergent were observed in pristine regions and cyclical transmission compared to the microheterogeneity showed by isolates from mechanical transmission areas.

Casos de malária importada, contraídos em região endêmica, mas diagnosticados em um país não-endêmico, são um evento raro mas com desfecho potencialmente fatal. Nosso objetivo foi investigar se a análise de genomas mitocondriais permite inferir a origem geográfica de casos importados de malária vivax diagnosticados nos EUA, comparando os resultados com aqueles obtidos por análise de DNA microssatélite. Foi sequenciado o genoma mitocondrial completo de 63 amostras de P. vivax provenientes de infecções importadas dos EUA, além de 7 amostras do Brasil e 6 do Panamá. A rede de haplótipos com DNA mitocondrial foi construída com 412 sequências e foi possível classificar com precisão a origem geográfica presumida dos isolados da América do Sul, Coréia, Sudeste Asiático e Melanésia, porém os isolados do Sul da Ásia, América Central e África não puderam ser classificados geograficamente. A análise bayesiana realizada com a tipagem de marcadores de microssatélites não apresentou sucesso quanto à classificação geográfica dos isolados de P. vivax. / Cases of imported malaria, infection acquired in an endemic region, but diagnosed in a non-endemic country, are rare but can lead to a fatal outcome. Our objectives were investigate whether the analysis of mitochondrial genomes allows inferring the geographic origin of isolates of P. vivax derived from cases of imported malaria diagnosed in the USA, and compare the performance of mitochondrial genome and DNA microsatellite analysis. We sequenced full mitochondrial genomes from 63 P. vivax isolates collected at the USA from imported infections, and 7 samples from Brazil and 6 Panama. A network of mitochondrial DNA haplotypes was built with 412 genomic sequences and were able to classify accurately isolates from South America, Korea, Southeast Asia and Melanesia according to their presumed geographic origin, but failed to do so with samples from South Asia, Central America and Africa. The Bayesian analysis performed by typing microsatellite markers showed no success on the classification of geographical isolates of P. vivax.

京都大学 / 新制・論文博士 / 博士(農学) / 乙第13412号 / 論農博第2895号 / 新制||農||1085(附属図書館) / 学位論文||R3||N5322(農学部図書室) / 京都大学大学院農学研究科農学専攻 / (主査)教授 冨永 達, 教授 土井 元章, 教授 那須田 周平 / 学位規則第4条第2項該当 / Doctor of Agricultural Science / Kyoto University / DFAM

Fusarium head blight (FHB) is one of the most destructive diseases of common wheat (Triticum eastivum L.) worldwide. Resistance to FHB is an ideal trait for which molecular marker assisted selection (MAS) would facilitate breeding and cultivar development efforts. Validation of quantitative trait loci (QTL) conferring FHB resistance is a prerequisite for MAS. This study was conducted to validate and evaluate the effect of two major QTL, previously reported on chromosomes 3BS and 5AS, on multiple FHB resistance components in two independent studies, one involving a mapping population derived from a cross between a known resistance source W14 and a susceptible soft red winter (SRW) wheat cultivar Pioneer2684, and the other involving seventy adapted SRW wheat lines. The first study confirmed that the 3BS and 5AS QTL were significantly associated with FHB resistance and further indicated that the 3BS QTL has a larger effect on three FHB resistance components (type II and III resistance and resistance to Fusarium Damaged kernels) evaluated in greenhouse experiments, while the 5AS QTL has a larger effect on type I resistance evaluated in a field experiment. Six simple sequence repeat (SSR) and two sequence targeted site (STS) markers associated with FHB resistance in the two QTL regions identified in the first experiment were then used to characterize FHB QTL marker haplotypes and their effect on FHB resistance in seventy wheat genotypes. Five main haplotype groups (1-5) were characterized among the elite lines on the basis of allelic differences of four marker loci linked to the 3BS QTL and two marker loci linked to the 5AS QTL. Haplotype group 5 was comprised of marker allele combinations of both 3BS and 5AS QTL and elite lines with this haplotype have improved type I and type II resistance compared to the other haplotypes. This again validated the presence of QTL on chromosomes 3BS and 5AS, and illustrated the utility of SSR and STS markers in the two QTL regions in selection of FHB resistance in elite backgrounds. Four favorable marker alleles including two (Xbarc133 and XSTS3B142) on 3BS and two (Xbarc117 and Xbarc056) on 5AS are recommended for MAS of the two QTL for improved FHB resistance in wheat. Wheat lines having favorable marker alleles identified in the current study will provide breeding programs with a source of unique and adapted FHB resistant parents and some of the lines also may have potential for release as cultivars. / Ph. D.

A despeito da importância comercial da cultura do maracujá-amarelo para o Brasil, estudos genéticos e de melhoramento são bastante escassos. Neste trabalho, foi conduzido um experimento visando construir mapas de ligação e mapear QTL para caracteres relacionados à produção e qualidade de frutos. Foi utilizada uma população composta por 160 indivíduos, proveniente do cruzamento de duas plantas dos acessos IAPAR-06 e IAPAR-123. Devido à alogamia da espécie, optou-se pela abordagem de mapeamento conhecida por "pseudocruzamento teste" para a construção de dois mapas de ligação, sendo um para cada genitor, utilizando marcadores moleculares do tipo AFLP, com segregação 1:1 e 3:1. No mapa do acesso IAPAR-06 foram obtidos 10 grupos de ligação, enquanto que no mapa do IAPAR-123, nove grupos foram obtidos. Os locos bi-parentais foram alocados como marcas acessórias nos mapas de ligação e serviram para o estabelecimento da homologia entre os grupos de ligação de cada genitor. Um total de oito grupos de ligação foi alinhado com base nesses locos. Em paralelo, procedeu-se a avaliação fenotípica de 100 indivíduos dessa mesma população, os quais foram avaliados em campo, na primeira safra de produção da cultura, para uma série de caracteres, quais sejam: velocidade de crescimento; produção total; número total de frutos; peso médio de frutos; comprimento e largura média de frutos; porcentagem de polpa; teor de sólidos solúveis totais; e formato médio de frutos. As análises genético-estatísticas dos dados fenotípicos indicaram que a população é amplamente variável para os caracteres avaliados, com exceção da velocidade de crescimento, apresentando coeficientes de herdabilidade entre médias que variaram de 52,6 a 83,0%. O mapeamento de QTL, utilizando o método de mapeamento por intervalo composto, identificou várias regiões associadas ao controle desses caracteres nos mapas individuais. Foram mapeados QTL para todos os caracteres que apresentaram variabilidade genética na população. A proporção da variação fenotípica explicada pelos QTL detectados variou de 4,6 até 21,8%. / Although the yellow passion-fruit plays an important commercial role in Brazil, breeding and genetics studies are insipient. The present study was conducted aiming the construction of linkage maps and mapping of QTL for traits related to yield and fruit quality. It was used a population, composed of 160 individuals, derived from a cross from two plants of the IAPAR-06 and IAPAR-123 accessions. Since this is an allogamous species, the pseudo testcross mapping strategy was used for the construction of two linkage maps, one for each parent, using AFLP markers segregating in 1:1 and 3:1 configuration. The IAPAR-06 map was composed of 10 linkage groups, while in IAPAR-123 map, nine linkage groups were obtained. The bi-parental loci were located as accessory markers and served to establish the homology between the groups of each parent. Eight linkage groups were aligned using these markers. For the phenotypic evaluation, 100 individuals were evaluated in the field, during the first harvest of the culture, for various traits, including: growth increment, total yield, total number of fruits, average fruit weight, average fruit length, average fruit width, percentage of pulp, soluble solids content and fruit shape. The results of the phenotypic data indicated that the population had a wide genetic variance for all the traits, with the exception of growth increment, presenting broad-sense heritability varying from 52.6% to 83.0%. The analysis of QTL, using the composite interval method, has mapped various regions associated with the traits evaluated in both maps. It was mapped QTL for all the traits that exhibited genetic variation. The proportion of the phenotypic variation explained by the QTL identified ranged from 4.6 to 21.8%.

Os marcadores microssatélites por serem de natureza codominante e multialélica tornam-se ideais para a caracterização, mapeamento e seleção assistida. Os locos microssatélites disponíveis têm sido utilizados efetivamente na caracterização de acessos de bananeira (Musa), porém muitos deles têm se mostrado com baixa eficiência de amplificação. Neste sentido, os objetivos deste trabalho foram: desenvolver novos locos de microssatélites, caracterizar 224 acessos de bananeira do banco de germoplasma da Embrapa em relação à ploidia e constituição genômica empregando abordagens moleculares e estabelecer relações genéticas entre eles. Foram desenvolvidos 50 novos locos de microssatélites dos quais 44 foram polimórficos entre os acessos analisados. Para a caracterização da ploidia dos acessos foi utilizado a citometria de fluxo, e para a definição da composição genômica foram empregados polimorfismo nos genes ribossomiais nucleares (rDNA) e 16 locos microssatélites, que também serviram para estabelecer as relações genéticas entre os acessos. Além disso, foram sequenciados regiões do rDNA de 17 acessos de bananeira para identificação de polimorfismo de base única (SNP) associados ao genoma A e B. A constituição genômica e/ou ploidia determinada por abordagens moleculares confirmou a classificação previamente estabelecida por descritores morfológicos, com exceção dos acessos Marmelo, Pitogo e Pisang Nangka. As regiões do rDNA sequenciadas permitiram identificar 17 SNPs específicos para M. balbisiana, que poderão ser utilizados para classificar acessos quanto a constituição genômica. A análise de agrupamento derivada da genotipagem dos acessos com 16 locos de microssatélites subdividiu os genótipos de acordo com a ploidia, constituição genômica e subgrupos de bananeira, enquanto que a abordagem empregando o modelo bayesiano corroborou de forma geral com essa categorização. Foi detectada uma ampla heterogeneidade nos acessos diplóides, onde poucos acessos triplóides tiveram sua ancestralidade definida nos grupos dos acessos diplóides. / Microsatellite markers are co-dominant and multiallelic, and the method of choice for genetic characterization, mapping, and assisted selection breeeding. Microsatellite loci have been effectively used to characterize banana (Musa) accessions, but in some cases they have shown low amplification efficiency. Thus, the objectives of this work were to develop new microsatellite loci; to characterize 224 accessions from the Embrapa germplasm collection of banana for ploidy and genomic constitution, as well as to establish genetic relationships among the accessions. Fifty new microsatellite loci were developed, from which 44 were polimorphic. Flow citometry was employed to characterize ploidy, while genomic constitution was determined by polymorphism at the nuclear ribosomal gene (rDNA) and 16 microsatellite loci, also used to establish the genetic relationship among the accessions. Additionally, rDNA regions were sequenced from 17 banana accessions to identify specific single nucleotide polymorphism (SNP) associated with the A or B genome. The genomic constitution and/or ploidy determined by various molecular approaches confirmed the previous classification established by morphological descriptors, except for Marmelo, Pitogo and Pisang Nangka. The sequenced rDNA regions allowed to identify 17 SNPs specific for M. balbisiana, which could be used to classify accessions according to genomic composition. Clustering analysis derived from accession genotyping using 16 microsatellite loci subdivided genotypes according to ploidy, genomic composition and banana subgroups, while the approach using a Bayesian model corroborated in general terms this categorization. A large heterogeneity of the diploid accessions was detected, which restrict the definition of ancestrality of triploid accessions in the diploid accession groups.

Uma espécie relevante como fonte alimentícia para a população mundial, principalmente para os países subdesenvolvidos e emergentes, é a mandioca (Manihot esculenta Crantz). Primariamente, a mandioca é fornecedora de energia a partir do amido acumulado em suas raízes, mas é importante destacar também a presença dos carotenoides com atividade pró-vitamínica A. Nesse contexto, o objetivo deste estudo foi caracterizar a diversidade agronômica, bioquímica e molecular de acessos de mandioca coletados nas regiões Amazônica, Centro-Oeste, Sudeste e Sul do Brasil, selecionados a partir da coloração de suas raízes. A caracterização foi realizada a partir de 81 genótipos (33 da região Amazônica, 24 da Centro-Oeste, 18 da Sudeste, quatro da Sul e duas variedade comerciais) utilizando 16 caracteres agronômicos, cinco bioquímicos e 17 locos microssatélites. Para os caracteres agronômicos e bioquímicos foram realizadas análises de variância, teste de Tukey e distância de Mahalanobis considerando a região de origem dos genótipos. Para os marcadores microssatélites foram estimados os índices de diversidade genética, tais como o número de alelos por loco (A), porcentagem de locos polimórficos (P), heterozigosidades observada (Ho) e esperada (He), bem como a estruturação da diversidade entre e dentro de regiões (HS, HT, DST e GST). Foram também obtidas a análise de agrupamento a partir da distância de Nei e método de Neighbor-Joining, e a estruturação dos genótipos a partir da análise Bayesiana. Além disso, as distâncias genéticas obtidas para os caracteres avaliados foram correlacionadas entre si pelo teste de Mantel. Os resultados obtidos apontam uma alta variabilidade para todos os caracteres avaliados. Para a maioria dos caracteres agronômicos houve variação significativa entre e dentro de regiões. Todos os caracteres bioquímicos apresentaram variação altamente significativa entre e dentro de regiões. Levando-se em consideração os caracteres agronômicos ocorreu uma tendência de agrupamento dos genótipos da Amazônia, sendo que as características que mais determinaram a distribuição dos genótipos foram a altura total da planta, a altura da primeira ramificação, a relação entre essas duas características e o índice de colheita. Para os caracteres bioquímicos ocorreu um agrupamento dos genótipos amazônicos em função da alta concentração de compostos cianogênicos e carotenoides totais. Os marcadores moleculares indicaram alta variabilidade genética (A = 3,58; P = 100%; Ho = 0,535; He = 0,642), sendo que a maior parte da diversidade (HT = 0,651) encontra-se distribuída dentro de regiões (HS = 0,643). A partir da análise de agrupamento foi verificada uma tendência de estruturação genética, com o agrupamento da maioria dos genótipos coletados na região Amazônica. A análise Bayesiana indicou a formação de dois grupos, sendo que um dos grupos englobou a maior parte dos genótipos da Amazônia, em concordância com a análise de agrupamento. As correlações entre as distâncias obtidas para todos os caracteres não foram significativas. Considerando o teor de compostos cianogênicos dos genótipos e a distância entre eles a partir dos marcadores microssatélites, foi constada uma tendência de separação entre os genótipos considerados bravos e os mansos. A partir dos resultados obtidos verificou-se que os genótipos da região Amazônica possuem características específicas que permitem o seu agrupamento. Essa tendência de agrupamento ocorre possivelmente devido a um processo de seleção diferenciado a que estes genótipos foram submetidos durante o processo de domesticação da espécie. / A relevant species as food source for the world population, especially for the underdeveloped and emerging countries, is cassava (Manihot esculenta Crantz). Primarily, cassava is a provider of energy from starch accumulated in its roots, but the presence of carotenoids with pro-vitamin A activity is also important. In this context, the objective of this study was to characterize the agronomic, biochemical and molecular diversity of cassava accessions selected for the color of their roots, collected in the Amazon, Central-West, Southeast and South regions in Brazil. The characterization was performed from 81 genotypes (33 from the Amazon region, 24 from Central-West, 18 from Southeast, four from the South and two commercial varieties) using 16 agronomic traits, five biochemical characters and 17 microsatellite loci. For biochemical and agronomic traits analyses of variance, Tukey test and Mahalanobis distances were obtained considering the region of origin of the genotypes. For microsatellite markers, indices of genetic diversity were estimated, such as number of alleles per locus (A), percentage of polymorphic loci (P), observed (Ho) and expected (He) heterozygosities, as well as the structuring of diversity between and within groups (HS, HT, DST and GST). Also, a cluster analysis was obtained using the Neighbor-Joining method and Nei´s genetic distance, as well as a Bayesian analysis. Moreover, the distances obtained for all characters were correlated using the Mantel test. Results indicate high variability for all characters evaluated. For most of agronomic traits there was significant variation between and within regions. All biochemical characters showed highly significant variation between and within regions. Taking into account the agronomic characters, there was a tendency for the Amazon genotypes to be grouped together, where the characteristics total plant height, height of the first branch, relationship between these characteristics and harvest index were those that mostly determined the genotypes distribution. For the biochemical characters, the Amazon genotypes were also grouped due to the high concentration of cyanogenic compounds and total carotenoids. Molecular markers indicated high genetic variability (A = 3.58; P = 100%; Ho = 0.535; He = 0.642), while most of total diversity (HT = 0.651) was found within regions (HS = 0.643). In the cluster analysis a tendency was observed towards a structured genetic variability, with the grouping of most genotypes collected in the Amazon region. The Bayesian analysis separated the genotypes in two groups, with one of the groups including most of the Amazon genotypes, in accordance with the cluster analysis. The correlations between the distances obtained for all characters were not significant. Considering the genotypes content of cyanogenic compounds and the distance between them from the microsatellite markers, a tendency was revealed for the separation between bitter and sweet genotypes. From the results obtained we may conclude that the Amazon genotypes have specific characteristics that allow its separate grouping. This tendency occurs possibly because of a differentiated selection process that these genotypes were submitted during the domestication of the species process.