Accessibility / Report Error

Population genetic analysis of Colombian Trypanosoma cruzi isolates revealed by enzyme electrophoretic profiles


Although Colombia presents an enormous biological diversity, few studies have been conducted on the population genetics of Trypanosoma cruzi. This study was carried out with 23 Colombian stocks of this protozoa analyzed for 13 isoenzymatic loci. The Hardy-Weinberg equilibrium, the genetic diversity and heterogeneity, the genetic relationships and the possible spatial structure of these 23 Colombian stocks of T. cruzi were estimated. The majority of results obtained are in agreement with a clonal population structure. Nevertheless, two aspects expected in a clonal structure were not discovered in the Colombian T. cruzi stocks. There was an absence of given zymodemes over-represented from a geographical point of view and the presumed temporal stabilizing selective phenomena was not observed either in the Colombian stocks sampled several times through the years of the study. Some hypotheses are discussed in order to explain the results found.

Trypanosoma cruzi; population genetics; spatial autocorrelation; Colombia

Population Genetic Analysis of Colombian Trypanosoma cruzi Isolates Revealed by Enzyme Electrophoretic Profiles

Vol. 96(1): 31-51, January, 2001

Manuel Ruiz-Garcia/ + + Corresponding author. Fax: +57-1-285.0503. E-mail: Received 8 December 1999 Accepted 15 August 2000 , Marleny Montilla*, Sebastian Nicholls*, Diana Alvarez

Unidad de Genética (Genética de Poblaciones-Biología Evolutiva), Departamento de Biologia, Facultad de Ciencias, Pontifica Universidad Javerina, Cra 7A No. 43-82, Bogotá DC, Colombia *Instituto Nacional de Salud, Laboratorio de Parasitología, Bogotá DC, Colombia

Although Colombia presents an enormous biological diversity, few studies have been conducted on the population genetics of Trypanosoma cruzi. This study was carried out with 23 Colombian stocks of this protozoa analyzed for 13 isoenzymatic loci. The Hardy-Weinberg equilibrium, the genetic diversity and heterogeneity, the genetic relationships and the possible spatial structure of these 23 Colombian stocks of T. cruzi were estimated. The majority of results obtained are in agreement with a clonal population structure. Nevertheless, two aspects expected in a clonal structure were not discovered in the Colombian T. cruzi stocks. There was an absence of given zymodemes over-represented from a geographical point of view and the presumed temporal stabilizing selective phenomena was not observed either in the Colombian stocks sampled several times through the years of the study. Some hypotheses are discussed in order to explain the results found.

Key words: Trypanosoma cruzi - population genetics - spatial autocorrelation - Colombia

In the present study, we show an extensive population genetics analysis, where the genetic heterogeneity and the spatial structure of Trypanosoma (Schizotrypanum) cruzi is analyzed throughout Colombia. This kind of study is of clinical importance since the extensive genetic polymorphism and heterogeneity of isolates may be related to the diverse pathology of the disease and its immunity, throughout the geographical distribution of the clonets (Dvorak 1984, Cosenza & Kroeger 1991). Although well kown in other Latin American countries (Brazil, Bolivia, Chile, for example), some population and spatial patterns of the T. cruzi clonets in Colombia are not very well known. The only previous studies reported for this country were those from Saravia et al. (1987), where 47 stocks were analyzed coming from the Piedmont region (Meta and Casanare departments); the inclusion of five Colombian stocks (MC50, MC52, MC53, MC60 and MC61) belonging exclusively to one zymodeme (19) in the study of Tibayrenc and Ayala (1988) and, more recently, the work of Márquez et al. (1998), where 28 new stocks coming from the Sucre, Antioquia and Tolima departments were analyzed. In the first work, 12 different zymodemes were detected (12/47 = 25.5%). In the second, the genetic variability shown in Colombia was lower, because only one zymodeme was detected from the five stocks analyzed (1/5 = 20%). In the third one, the amount of diverse zymodemes increased notably because 21 new zymodemes were detected out 28 stocks analyzed (21/28 = 75%). However, these stocks came from sylvatic origin and did not include stocks representing the domestic cycle of this parasite, and only represented a limited geographic fraction of Colombia. In our analysis, we included 23 stocks coming from seven different departments of Colombia (North Santander, Cesar, Antioquia, Casanare, Cundinamarca, Tolima, and Boyacá), and 20 new zymodemes were detected (20/23 = 87%). This evidences that the previous studies underrepresented the real genetic diversity in this neotropical country. It is remarkable that 18% of the Colombian territory (about 200,000 km2) containing three million people is placed in the endemic area of distribution of T. cruzi, with almost one million people infected (3% of the total population of this country). The main population genetics results were in agreement with the hypothesis of a clonal structure, although recombination can be induced in vitro in several bacteria and other microorganisms, such as Escherichia coli (Ochman & Selander 1984). However, some spatial results showed the presence of various relevant geographic patterns, which shows the lack of one, or some, over-represented zymodeme(s) over broad geographical areas. However, we show that the determination of spatial structure strongly depends on the statistic techniques employed. Additionally, the presumed temporal stabilizing selective phenomena are not observed either in the Colombian stocks sampled several times throught the period studied.

Populations studied - Samples from people infected with T. cruzi, and animal samples from wild reservoirs, as well as vectors, were collected from 1985 to 1995 in North Santander, Cesar, Antioquia, Casanare, Cundinamarca, Tolima, and Boyacá, covering a significant fraction of the endemic geographical area of Colombia where this protozoan is present. The hosts and vectors, in which T. cruzi was obtained, were humans, Rhodnius prolixus, Panstrongylus geniculatus and the marsupial Didelphis marsupialis. Upon culture, isolation and characterization, 23 different stocks of T. cruzi were established. The areas of Colombia examined correspond to endemic areas of T. cruzi covering a broad range of altitudes, temperature and environmental conditions. Two reference stocks were additionally surveyed with regard to the Colombian stocks analyzed. One of these stocks comes from Bolivia (SO34), isolated from Triatoma infestans, and represents the Z1 zymodeme, predominant in Brazil and with a wild origin (Miles et al. 1984). The second one comes from Chile (MN), isolated from humans, and belongs to the Z2 zymodeme of typical domestic cycle. For overall comparison purposes, two samples of Crithidia sp. were analyzed along with the 23 Colombian stocks of T. cruzi. The stocks were isolated using a modified NNN culture.

Markers used - For the isoenzymatic analysis, 106 to 108 cells/ml from cultured stocks of T. cruzi were obtained by centrifugation at 1,000 g for 10 min at 4oC. Upon resuspension in sterile normal saline solution, the cell pellet was kept at -70oC. Electrophoresis in cellulose acetate was done according to the procedures of Tibayrenc et al. (1985). Thirteen different enzymatic markers were analyzed: glucose-6-phosphate isomerase (Gpi), malate dehydrogenase (Mdh), isocitrate dehydrogenase (Idh), phosphoglucomutase (Pgm), glutamate dehydrogenase NAD+ (Gdh-Nad+), glutamate dehydrogenase NADP+ (Gdh-Nadp+), malic enzyme 1 (Me-1), malic enzyme 2 (Me-2), glucose-6-phosphate dehydrogenase (G6pdh), peptidase 1 (Pep-1), peptidase 2 (Pep-2), 6-phosphogluconate dehydrogenase (6pgdh) and aspartate aminotranspherase (Got). The first 12 of these isoenzymes had also been analyzed by Tibayrenc and Ayala (1988) in their description of 43 American zymodemes, being the only difference with our set of polymorphic markers the use of Got instead of leucine aminopeptidase (Lap). In their study, they had also included aconitase and adenylate kinase, both of which resulted monomorphic and were not included in the profiles of the 43 zymodemes they defined. The comparison between their zymodemes and those presented in this analysis will therefore be possible to a large extent.

Population genetics procedures

Hardy-Weinberg equilibrium - To analyze the possible existence of Hardy-Weinberg equilibrium individually and collectively in the T. cruzi stocks analyzed, an unbiased estimator (f) based on the proportion of homozygotic alleles, which can be used as a measure of the coefficient of endogamy (Robertson & Hill 1984), was used. In this analysis, we assumed the hypothesis of diploidy in T. cruzi.

Heterozygosity and heterogeneity - The presence of stochastic or selective differential events could be detected in a given population if the heterogeneity of the H statistic (= expected heterozygosity) was too large. For this reason, this statistic was calculated for each stock analyzed. Additionaly, the GST statistic was obtained to analyze the degree of heterogeneity among the stocks studied.

Genetic distances - To analyze the genetic relationships among the Trypanosoma stocks, diverse genetic distance matrices were calculated (Nei's, Cavalli-Sforza & Edwards and Prevosti). Various dendrograms were built using diverse algorithms. There were the Single (Single-linkage clustering) and the WPGMA (Weighted Arithmetic Pair-Group Method) algorithms. The second procedure was applied by using the method of Pamilo (1990). The method of "neighbor-joining" was additionally applied (Saitou & Nei 1987) which yields trees without origin root, and for this we applied the midpoint rooting method (Farris 1972) to produce trees with direction. The robustness of the genetic dendrograms obtained was assessed by two methods. The first was the Felsenstein's bootstrap test (1985). The second method of assessment was the obtainment of the cophenetic correlation coefficients (Sneath & Sokal 1973), which was applied to all the dendrograms generated, although we only present those with the highest values. The statistical significance of those coefficients was obtained applying an approximate Mantel t test, and using a Monte Carlo simulation with 1,000 permutations.

Spatial autocorrelation analysis - We carried out some autocorrelation analyzes to study the spatial structure of the allozymic markers. Having revealed an extensive genetic heterogeneity for the 13 loci studied in the 23 Colombian stocks of T. cruzi, we searched for a spatial structure of the populations of this organism in Colombia, by using two different analyses: (1) the calculation of two indices of spatial autocorrelation, the index I of Moran, and the c coefficient of Geary (Sokal & Oden 1978a,b). Two different analyses of spatial autocorrelation were carried out with the indices mentioned. The different alleles of each loci were used as allelic frequencies, and the less frequent of them were excluded to break the relationship of lineal dependence of all the alleles of a given locus. Fifty one variables were studied in this case. With this first spatial autocorrelation analysis, two different distance classes were elected. In the first, seven distance classes (DC) were defined, with the number of stock pair comparisons identical for each DC: 1 DC = 0-97 km, 2 DC = 97-166 km, 3 DC = 166-213 km, 4 DC = 213-295 km, 5 DC = 295-351 km, 6 DC = 351-423 km, 7 DC = 423-580 km. The number of stock pair comparisons for each DC was 15. In the second analysis, 8 DC were elected, where the size of the DC was constant independently of the stock pair comparison number: 1 DC = 0-73 km, 2 DC = 73-145 km, 3 DC = 145-218 km, 4 DC = 218-290 km, 5 DC = 290-363 km, 6 DC = 363-435 km, 7 DC = 435-508 km, 8 DC = 508-580 km. Both analyses were carried out to analyze if any change in the spatial pa-rameters could affect the results obtained. On the other hand, the various genotypes for each locus were used as unordered multistate characters. For example, if for MDH three genotypes were found (1/1), (2/2) and (3/3) in the 23 different stocks studied, therefore these would represent the states 1, 2 and 3, respectively. If for PGM four different genotypes were found, (1/1), (3/4), (5/7) and (5/8), the respective multistate values would be 1, 2, 3 and 4. An analysis of spatial autocorrelation was applied, observing the various genotypes. Thirteen variables were thus defined, as were seven DC, for constant geographical sizes, and disimilar number of comparisons per DC: 1DC = 0-83 km, 2DC = 83-166 km, 3DC = 166-249 km, 4DC = 249-332 km, 5DC =332-415 km, 6DC = 415-497 km, and 7DC = 497-580 km. It is interesting to analyze whether using allelic frequencies, or multistate genotypic forms, similar spatial results are obtained. A contrasting result would indicate the differential importance in the choice of representative values for the alleles and for the genotypes. In addition, for each collection of data, a single autocorrelation coefficient was obtained for each genetic variable studied. In this case, the point pairs were weighted as the inverse square separation distance between the stocks.

The kind of correlograms obtained may suggest the type(s) of evolutionary mechanism(s) leading to a given spatial structure (Sokal & Wartenberg 1983, Sokal et al. 1987, 1989, Sokal & Jacquez 1991, Ruiz-Garcia 1994a, b, Ruiz-Garcia & Jordana 1997, 2000, Ruiz-García & Klein 1997).

Analysis of similarity between correlograms - To determine the similarity between the cor-relograms, the average Manhattan distance matrices between the coefficients of autocorrelation estimated for variable pairs of correlograms were calculated, both for allele frequencies (7 and 8 DC) and for genotypes. This analysis can determine whether each of the genetic variables studied was subject to the same spatial evolutionary event, or whether they were under pressure from different spatial evolutionary agents. Sokal and Wartenberg (1983) and Sokal et al. (1989) showed, by means of simulation studies, that correlogram pairs generated by the same evolutionary spatial processes have Manhattan distances smaller than 0.1 in the case of Moran's I index, and about 0.15 for the Geary's c coefficient. In order to visualize the spatial relationship between pairs of variables, the algorithm UPGMA was applied to the Manhattan distance matrix between the correlograms corresponding to the variables for each spatial autocorrelation analysis performed.

The 20 new zymodemes detected in the T. cruzi stocks of Colombia are different from those reported in this country by Saravia et al. (1987), Tibayrenc and Ayala (1988) and Márquez et al. (1998) (see Table I). Otherwise, some caution has to be taken in account with direct comparisons with already published data since techniques were not performed by the same operators. With this in mind, a detailed comparison for each locus, contrasting our results with those reported by Tibayrenc and Ayala (1988) and Márquez et al. (1998), is as follows:

Table I

Gpi: the homozygote (5/5) had been detected in Colombia by Tibayrenc and Ayala (1988). Lately, Márquez et al. (1998) detected the genotypes 8/8; 4/7; 4/4; 3/9 and 9/9. In this study we show four different genotypes, of which three had been reported in other locations of Latin America (3/3; 4/4; 6/6) and the remainder (7/7) is new.

Mdh: the homozygous (2/2) had previously been reported in Colombia and was in fact present in many of the analyzed stocks. The homozygous form (3/3) is however new, both in Colombia and in Latin America.

Pgm: three genotypes unknown in Colombia and in Latin America were detected (3/4; 5/7; 5/8). The homozygotes (1/1; 3/3; 4/4; 5/5) and the heterozygote (2/3) were already known in Colombia.

Idh: two Colombian unknown genotypes were found (2/2; 3/3), being (3/3) also new in Latin America. The form (1/1) had previously been reported in this country.

Gdh-Nad+: five different genotypes were present; (3/3) had already been reported in Colombia. Genotypes (4/4 and 5/5) are reported for the first time in Latin America. A new probable genotype of null alleles was detected.

Gdh-Nadp+: three genotypes were detected, of which only (2/2) had been detected in Colombia, whereas (3/3) had been found in Latin America. A third genotype probably represents null alleles.

Me-1: three genotypes were found; (2/2) had already been reported in Colombia. Genotypes (2/2) and (3/3) had been found in other parts of Latin America. A compound genotype was detected, probably of null alleles.

Me-2: three genotypes were found (2/2; 3/3; 5/5) all of them unknown in Colombia, where only genotype (4/4) had been reported. In Latin America genotypes (3/3 and 5/5) had been reported, but genotype (2/2) is new.

G6pdh: four diverse genotypes were detected in Colombia (0/0; 2/2; 3/3; 4/4). Only genotypes (5/5, 3/3 and 2/2) had previously been reported in this country, differently from genotypes 0/0 and 4/4, which are new in Colombia. All of them had already been detected in other parts of Latin America, except for that composed of null alleles.

Pep-1: four different genotypes were detected (2/2; 4/4; 5/5; 6/6). Genotypes (1/1; 5/5; 4/4; 2/2) had previously been reported in Colombia. Genotype 6/6 reported here is new. All of them had already been reported in other places of Latin America. Curiously, Márquez et al. (1998) detected an important number of heterozygote genotypes for this locus (2/7; 2/5; 1/3; 2/6; 1/7; 2/3) in Colombia, while we did not detect any.

Pep-2: genotypes (1/1; 2/2) were detected, as well as a null form. Of these, genotype (1/1) had previously been detected in Colombia.

6pgdh: genotype (4/4) had been reported in the previous study by Tibayrenc and Ayala (1988). In contrast, we report new genotypes (0/0; 5/5; 6/6) in Colombia, although previously reported in Latin America, except for the null genotype.

Hardy-Weinberg equilibrium - The Hardy-Weinberg equilibrium analysis showed the inexistence of this equilibrium at all levels. In this case, by using the f of Robertson and Hill (1984), an excess of homozygosis was found for the total population (f = 0.1910; c2 = 5.471; 1 df; P <0.05; see Table II), although applying the F of Wright, a non significant excess of heterozygosis was obtained (F = -0.3158; c2 = 14.958, 21 df; P =0.825). Márquez et al. (1998) found an overall observed hetero-zygosity, which was not statistically different from the Hardy-Weinberg proportions. However, the overall and the individual absence of Hardy-Weinberg equilibrium of this magnitude in our data is typical of a clonal structure.

Table II

Levels of individual diversity and hetero-geneity in the T. cruzi stocks surveyed in Colombia - In this study the proportion of different zy-modemes for Colombia is noteworthy higher (87 %) than the values obtained by previous authors. However, the mean heterozygosity (H) found in this work (around H = 0.04) is substantially lower (ten times lesser) than the value reported by Márquez et al. (1998) (H = 0.40). Our value is more similar to that reported by Tibayrenc and Ayala (1988) (H = 0.06). In addition, we found only one locus (PGM) in heterozygote condition out of the 13 loci analyzed (7.7 %). On the contrary, Márquez et al. (1998) found heterozygous genotypes for the loci Alat, Mpi, Pep, Gpi, Acon, Icd, Ldh, Asat and Pgm (in lower magnitude for this locus than the value reported here) out of the13 loci analyzed (9/13 = 69. 2 %).

Our estimate of genetic heterogeneity among the T. cruzi stocks analyzed in Colombia is one of the highest reported for whatever organism in the world (GST = 0.93) and it is strongly significant. For 33 Colombian clonets, Márquez et al. (1998) obtained a FST range of 0.33-0.67, also highly significant. Therefore, there is agreement between both works in this aspect.

Genetic distances - The genetic distances found between pairs of Colombian stocks of T. cruzi are enormous in many cases, suggesting an extensive degree of genetic diferentiation between pairs of stocks. Tibayrenc et al. (1986b) found for Latin America values of Nei's distances ranging from 0.017 to 2.015, with a mean value of 0.757 ± 0.478. Although the geographical span analyzed in Colombia is considerably smaller, the values found ranged from 0.0-0.147 to 1.833, with mean values according to those for all Latin America by Tibayrenc and Ayala (1988) (Table III). This means that the enormous divergence between the stocks studied in Colombia is of the same order of magnitude than that found by Tibayrenc and Ayala (1988) for 121 stocks covering all Latin America, including populations of USA. The results obtained applying the distance of Nei were also found using the other genetic distances calculated. The Chritidia stocks differed significantly from those of T. cruzi with Nei's distances of 2.545. Remarkably, the values between pairs of stocks of T. cruzi and Chrithidia were in some cases smaller than those between certain pairs of stocks of T. cruzi.

Table III

The most prominent dendrograms were observed when applying the algorithm WPGMA, and the distances of Prevosti and Cavalli-Sforza & Edwards. Both dendrograms (Fig. 1) clearly separated Crithidia from all T. cruzi stocks. The first T. cruzi stock which diverged was that of Chile (MN) representing the Z2 zymodeme of domestic cycle. Otherwise, the Bolivian stock (SO34) was more related to the Colombian stocks, showing that these stocks basically belong to the Z1 zymodemes. Inside of the Colombian array, those of Amalfi (Antioquia) and some of the Cundinamarca department (central Colombia) were the first which diverged. However, there is not a clearly defined number of clusters within the departments. These dendrograms highly differed from the Single analysis with the Nei's distance because of grouping together stocks 5 (Arboledas, North Santander) and 7 (Zetaquirá, Boyacá) with the cluster where are stocks 1 (El Zulia, North-Santander) and 8 (Guateque, Boyacá). The dendrogram generated with the method of neighbor-joining and the distance of Prevosti leads to a similar result, displaying it more clearly than the other methods the group comprised by stocks 5, 6, 7 and 12.

Figure 1

Spatial autocorrelation analysis - The interpretations based on the results derived from the spatial autocorrelation analysis, due to the use of genotypic multistates, or the use of the allelic frequencies, are remarkably different.

Genotypic multistates: by using Moran's I index, only G6pdh out of the 13 variables studied was significant (p = 0.005) with a clear circular clinal structure, showing a 1DC significantly positive, 4DC significantly negative, and the 5DC significantly positive again when seven DC were used. Got follows a similar spatial pattern, but it did not reach the level of statistical significance. Two other markers, Gdh-Nadp+ and 6pgdh, also stood close to the point of statistical significance, and displayed a strong decline in genetic similarity at 4DC and 3DC, respectively (Table IV). The average correlogram is clearly not significant. With the Geary's c coefficient, the same trends are observed more clearly. G6PDH showed a significant circular cline, as with the Moran's I index. In contrast to this, Gdh-Nad+ presented a significant case of differentiation at long distance (Sokal et al. 1989), with an extremely high value at 7DC (c = 3.70; p <0.0001). Gdh-Nadp+ also displayed a significant circular cline. The remaining variables did not display any kind of spatial structure. In general, the application of genotypic multistates did not reveal the presence of much significant spatial structure. This is reflected in the following results: both the percent of significant coefficients of autocorrelation, as well as the percent of significant global correlograms, did not differ from the 5% of type I error for any of the two indices applied (Moran and Geary: 8.79%; 8/91; c2 = 1.05, 1 df, NS; Moran: 7.69%; 1/13; NS; Geary: 23%; 3/13; NS, respectively). Neither the 13 variables analyzed showed significant values for a single autocorrelation coefficient without defining distance classes (Table V).

Table IV

Table V

The analysis of similarity for the correlograms derived from the matrices of distances of Manhattan for the Moran's I index is based on the percent of distances between the correlograms of pairs of variables with values smaller than 0.1. This percentage was 1.3%, which did not even reach the error type I of 5%. By applying genotypic multistates, it was not possible to reveal any degree of similarity in the spatial distribution of the 13 variables analyzed (Fig. 2A).

Figure 2

Allelic frequencies- The use of the allelic frequencies, in contrast, revealed a significant spatial structure of various kinds for the 51 variables analyzed for the 7 and 8 DC defined. The alleles that showed significant global correlograms with the Moran's I index and with 7 DC were 12 (12/51 = 23. 529 % of significant overall correlograms) and there was 10.9 % of significant individual autocorrelation coefficients (39/357), both percentages significant (Table VI): Got 7, Gdnad 5, G6pdh 4 and Pep1 5 with a clear genetic differentiation at long distance; Gdh-Nadp+0 presents a typical symmetrical circular cline with 4DC being significantly negative, and 1DC and 7DC the most positive of all seen; Me1-0, Me2-1, G6pdh-1, Pep2-3 and 6pgdh 1 with a significant global structure of regional genetic patch from 166 to 213 km; G6pdh 2 has the structure of a regional patch, of a considerably larger size than the other alleles with this structure (1DC significantly positive, and 6DC significantly negative) and G6pdh 3 with high genetic similarity in the first 213 km and strong genetic differences from 351 km. In addition, other nine alleles showed correlograms very near to the significance border (9/51 = 10.92%) Gpi 2, Got 0, Mdh 1, 6pgdh 0 and Pgm 1, with a near global structure of regional genetic patch from 166 to 213 km. The high similarity between the spatial patterns for the frequencies of those alleles at different loci may be an important evidence of a strong linkage disequilibrium between groups of alleles of different loci. Gnadp 3 showed a correlogram typical of a circular cline, Pep2 2 and Pep2 0 presented negative values for the first distance class but significant ones for the second class distance (97-166 km). With 8 DC defined the results were similar. In fact, the spatial genetic structure is larger and highly significant, as well. For the Moran I index, 31.4% (16/51) of the overall correlograms and 10.5% (43/408) of the individual autocorrelation coefficients were respectively significant. The alleles which showed significant correlograms were Gpi 2, Gpi 3, Got 0, Got 7, Mdh 1, Pgm 1, Gdnad 5, Gnadp 0, Me1 0, Me2 1, G6pdh 2, G6pdh 3, G6pdh 1, G6pdh 4 Pep2 3 and 6Pgdh 1. The vast majority was significant with 7 DC defined, or was significant borderline, as well.

Table VI

Applying the Geary's c coefficient, the situation was similar. With 7 DC defined, 27.4% of significant overall correlograms (14/51) and 15.4% of significant individual autocorrelation coefficients (55/357): Got 7, Gdnad 5, G6pdh 4 and Pep1 5 showed significant differentiation at long distance; Gdnadp 3, Gdnadp 0 and G6pdh 0 yielded circular clines; Me1 0, Me2 1, G6pdh 1, Pep1 2 Pep2 3 and 6pgdh 1 showed significant regional patches; G6pdh 3 showed striking genetic similarity until 213 km and a strong disimilarity between the stocks at 351 km, and G6pdh 2 showed a monotonic clinal divergence until 423 km, with the last DC not significant. Additionally, 13.7% of other correlograms were borderline significant (7/51; Gp1 2, Got 0, Mdh 2, Mdh 1, Pgm 5, Pgm 1 and G6pdh 0). With 8 DC, 33.3% (17/51) of overall correlograms and 12.5% (51/408) of individual autocorrelation coefficients were significant, respectively. The alleles showing significant correlograms were Gp1 2, Got 0, Got 2, Got 7, Mdh 1, Pgm 1, Gdnad 5, Gdnadp 0, Me1 0, Me2 1, 6pgdh 3, G6pdh 1, G6pdh 4, Pep1 5, Pep1 2, Pep2 3 and 6Pgdh 1. The spatial trends found were very similar to those obtained with the Moran's I index and with 7 DC. On the contrary, when only 6 DC were considered, with this coefficient, some spatial significant structures changed. Pgm 5 and Pep-1 5 showed significant global correlograms as regional patches, and differentiation at long distance, respectively. Furthermore, with 6 DC, 13 alleles showed significant global correlograms, and 19 alleles showed a tendency to spatial arrangements of no statistical significance.

There seems to be a slightly better chance to reveal spatial structures, or trends, with the use of the coefficient c of Geary than with the I of Moran. For instance, with 6 DC some of the alleles which were not detected with the Moran's I index were Got 4 (high similarity among neighbor stocks), Got 7 (differentiation at long distance), Got 0 (regional patch), Got 2 (circular cline), Pgm 8 (regional patch) and Gdh-Nadp+ 3 (circular cline). It is interesting to note that each allele of the Got locus yielded a different spatial trend. The arrays of alleles that presented strong linkage in their spatial structures were basically the same by using the Moran's I index and the Geary's c coefficient, and were the same independently of 6, 7 or 8 DC chosen. When no distance classes were considered, the individual autocorrelation coefficient for each allele consisted in 105 pair comparisons. For the Moran's I Index, the alleles Gdnad 2 (P = 0.048), Gdnadp 2 (P = 0.05), G6pdh 3 (P = 0.01) and 6Pgdh 5 (P = 0.008) were significant. This means that these were the alleles which displayed more significant spatial structures. For the Geary's c coefficient, the alleles Gdnad 2 (P = 0.014), Gdnadp 2 (P = 0.014), G6pdh 3 (P = 0.013) and 6Pgdh 5 (P = 0.008) were also significant (Table VII). Altogether, it can be stated that the amount of spatial structure found in Colombia for T. cruzi is noteworthy significant. Some of the overall results support this view. The percentage of significant coefficients of autocorrelation was superior to the error type I of 5%, both for the Moran's I and for the c of Geary for 6 DC (17.3%; 53/306; c2 = 23.4; 1 df; P <0.001), for 7 DC (Moran: 10.9%; Geary: 15.4%) and for 8 DC (Moran: 10.5%; Geary: 12.5%). A percent value above the error type I of 5% was also found for both indices related with the number of significant global correlograms for 6 DC (Moran: 23.5%; 12/51; c2 = 7.2; 1 df; P <0.01; Geary: 25.5%; 13/51; c2 = 8.3; 1 df; P <0.01), for 7 DC (Moran: 23.5%; Geary: 27.4%) and for 8 DC (Moran: 31.4%; Geary: 33.3%). Therefore, there is not doubt about the striking spatial structure among the Colombian T. cruzi stocks, when the allele frequencies are employed.

Table VII

It is of upmost importance to determine whether it is better to obtain the spatial autocorrelation by using the allelic frequencies or the genotypic multistates. We find the use of the allelic frequencies much more informative than that of genotypic states, since for the latter we have assigned the value of each state randomly, while ignoring the type of order and evolutionary transition that might have existed between the different genotypes. In contrast, with the results obtained when using the genotypic multistates, the analysis of correlogram similarity based on the allelic frequencies showed significant percent values of Manhattan distances. For instance, for the Moran's I Index, a percent of Manhattan distances smaller than 0.1 of 8.8% (113/1275) was obtained for 6 DC and 8.4% (107/1275) for 7 DC, being both significantly superior to the error type I of 5% (c2 = 14.6; 1 df; P < 0.001 and c2 = 11.7; 1 df; P < 0.01). There is, therefore, a significant fraction of variables with the same patterns of spatial arrangement. For 6 DC, and with the Moran's I coefficient, there was the presence of ten clusters of highly spatial related alleles (Fig. 2b): (1) Gpi 6 and Mdh 3; (2) Got 4, Got 7, Gdh-Nad+ 5, G6pdh 4, Pep-1 5 and Idh 2. This array was comprised by the variables that have spatial structure of differentiation at long distance. Except in a few cases, it was observed that most spatial clusters contained alleles from different loci, a manifestation of linkage disequilibrium; (3) Gpi 2, Got 0, Mdh 1, Pgm 1, Me-1 0, Me-2 1, 6pgdh 1, Pep-2 3, Gdh-Nad+ 4 and 6pgdh 0. These were the alleles which typically showed correlograms describing regional patches: (4) Gpi 4, Pgm 0, Gdh-Nad+ 0, Pgm 7, Pgm 3, Pgm 4, 6pgdh 5, and Pep-2 0; (5) Gdh-Nad+ 2 and Gdh-Nadp+ 2; (6) Me-2 and 6pgdh 6; (7) Gpi 3, Pep-1 6, Pep-2 1, Me-1 3 and Me-2 3; (8) Me-1 2 and Me-2 2; (9) Pgm 5 and Pep-1 2 ; and (10) Got 2 and G6pdh 0.

When 7 DC were used, eight clusters of highly spatial-related alleles were conformed with the Moran's I Index and nine with the Geary´s c coefficient. Some of these allele-spatial related clusters were slightly different among them (Fig. 2C,D).

In addition, we studied the similarity among the various distance classes to understand the contribution of each one to the spatial behaviour of the genotypes and allele frequencies studied (Fig. 3). For the analysis with 7DC, the most divergent distance class, simultaneously for genotypes and for allele frequencies, was the 7 DC. Therefore, this is the distance class with the most differentiated spatial behaviour, which is in agreement with certain isolation-by-distance structure.

Figure 3

Some of the results of the population genetics analysis presented here support the point of view of a clonal structure of the Colombian T. cruzi stocks proposed by Tibayrenc (1996) and Tibayrenc and Ayala (1988, 1991).

Clonal population structure

Hardy-Weinberg equilibrium - We found total absence of Hardy-Weinberg equilibrium in the stocks of the Colombian T. cruzi. This is a clear evidence of clonal structure. When the absence of equilibrium Hardy-Weinberg of this parasite is seen in the context of the genetic structure of its vectors and hosts an additional result is obtained, which indirectly supports the view of the clonal structure. Dujardin et al. (1987) studied the genetic structure of the vector T. infestans in Bolivia and demonstrated that it is in Hardy-Weinberg equilibrium. Tibayrenc et al. (1991) found in Hardy-Weinberg equilibrium the populations of humans exposed to various species of Trypanosoma, having studied several markers in more than 1,600 people in South America, Europe, Africa and French Polynesia. Previous studies of human populations in Colombia, by other authors and ourselves, have found these Colombian human populations to be in Hardy-Weinberg equilibrium for most markers used (Castillo & Ruiz-García 2000, for HLA DQa, LDLR, GYPA, HBGG, D7S8 and Gc; Jaramillo-Correa et al. 2000, for APO-E and ACE; Duran & Ruiz-García 2000 for VWA and TH01). That is to say, apparently the genetic structure of human populations did not condition the absence of Hardy-Weinberg equilibrium in the Colombian T. cruzi stocks. Perhaps, if some vector, or host, were not in H-W equilibrium, it could either reinforce the clonal structure, or hide the presence of a panmictic structure to a certain degree. There are indications within our data to support this view. The three stocks, with less genetic diversity (a value of zero), were obtained from R. prolixus, whereas the rest of the stocks, with some degree of identical genetic diversity were from humans and from D. marsupialis.

Tibayrenc and Ayala (1991) have put forward the view that the excess of homozygotes is not explained by the Wahlund effect. We also believe that the Wahlund effect is not a reasonable candidate to explain the evolution of the current genetic profiles found in Trypanosoma, although we do not share their explanation for it. The Wahlund effect could in fact be taking place in certain points of the distribution of a species, but not necessarily at every point within its range of distribution; this conjunction could however be the product of a clonal structure, unless the structure of the species is comprised by microdemes with high reproductive isolation (Ehrlich & Raven 1969), and the sampling procedure provided was not fine enough at the microgeographical level. The presence of loci fixed in heterozygosity is not sufficient to state that this might not be constituted by Wahlund effect. Negative Wahlund effect due to migration and in-breading among individuals from different genetic pools (Ruiz-García & Alvarez 2000), patterns with phylopatric females and migratory males (Chesser 1991a,b), overlaping of different generations with cross breading (Ennos 1985, Lopez-Alonso & Pascual-Reguera 1989) can generate patterns of excess of heterozygotes. This is not parsimonic enough, and is the reason why the Wahlund effect does not seem to be the cause for the observed patterns, is the possibility that simultaneous positive and negative Wahlund effect might be taking place, at such magnitude, for different loci of the same stocks.

Linkage disequilibrium - The existence of strong linkage disequilibrium seems to be an essential feature of the clonal structure. Tibayrenc et al. (1986a,b) applied the method of Ohta (1982) to detect linkage disequilibrium in subdivided populations and found high levels of disequilibrium for the stocks they analyzed, even higher than those found in self-fertilizing species such as Hordeum spontaneum. We have not analyzed here the linkage disequilibrium in particular, but have found a significant percentage of variables with identical correlograms, and are therefore affected by the same type of evolutionary events. The high degree of spatial structure similarity in many of the alleles analyzed for different loci could be due to the presence of a significant linkage disequilibrium. Ruiz-García and Klein (1997) demonstrated that the absence of gametic disequilibrium in populations of cats was associated to the non existence of similar correlograms between the loci analyzed. Our results are in agreement with those of Márquez et al. (1998) and Lewicka et al. (1995), who encountered considerable linkage disequilibrium to analyze 24 T. cruzi stocks in French Guiana. However, it is impossible to rule out the possibility of some residual sexuality, especially in the sylvatic cycles. Bogliolo et al. (1996) and Carrasco et al. (1996) have reported some possible cases of hibridization. The last authors analyzed 36 Amazonian T. cruzi stocks, and, apparently, found the Pgm locus in Hardy-Weinberg equilibrium, and some RAPD profiles seemed parentally homozygotes and its hybrids heterozygotes. As we show in Fig. 4, the Colombian T. cruzi are strongly related with the Bolivian SO34 T. cruzi, which represented the Z1 zymodeme of sylvatic origin, and not with some Chilean MN stocks representing the Z2 zymodeme of domestic origin. This result agrees with those obtained in the present study with 23 new Colombian T. cruzi stocks, and also with that reported by Márquez et al. (1998) who found that the major part of the Colombian stocks they studied were closely related to the Z1 zymodeme of sylvatic origin. This means that sylvatic zymodemes are the most frequent in domestic conditions in Colombia, and the probability of some genetic recombinant events could be more frequent due to this sylvatic origin than in the domestic zymodemes (Z2, Z3), which prevail in other Latin American countries (Chile, Argentina, part of Bolivia, etc.). In addition, this dendrogram puts forward that the 23 T. cruzi stocks surveyed here were not especially related to other forms of Trypanosoma, such as T. rangeli or T. cruzi marenkellei. This last taxa was represented by Trypanosoma sp. from Antioquia and conformed a different cluster.

Figure 4

Genetic heterogeneity - The mean genetic heterogeneity found in this study (GST = 0.934) is one of the highest values ever reported for any organism, the values for this statistic being commonly lower than 0.1. This is another symptomatic event of clonal structure. To illustrate this, the values found in various animal species are as follows: cats in Spain, Balearic Islands and Italy (GST = 0.03, 0.02 and 0.03, respectively, Ruiz-Garcia 1993, 1994a, Randi & Ragni 1991), Yanomamas (GST = 0.069, Weitkamp et al. 1972), Drosophila pseudoobscura (GST = 0.008 Alvarez et al. 2000), brown trout (GST = 0.041, Chakraborty et al. 1982), Canis latrans (GST = 0.08, Hamilton & Kennedy 1986). In mammals, two of the highest values reported for this statistic have been for Mus musculus in Europe (GST = 0.465, Britton-Davidian 1990), and Dipodomys ordii (GST = 0.674, Johnson & Selander 1971). Some of the values reported for plants have been H. spontaneum (GST = 0.25, Baum et al. 1997; GST = 0.284, Nevo et al. 1986), and Triticum turgidum (GST = 0.37, Nevo et al. 1988). None of these values get near that obtained for Trypanosoma. As observed in the study of Tibayrenc and Ayala (1988), the range of the genetic distances found is enormous, and some of the values are extremely high (more than 1.8 for the Nei's distance). This is another evidence of clonal structure. For example, Ayala (1975) found in Drosophila, for this distance, mean values of 0.581 for sinmorphic species, and 1.056 for alomorphic species. Some of the values we found are higher, though it does not mean that we have stocks from different species. The continuum values for these distances between pairs of stocks suggest that we face a species of exceedingly high heterogeneity due to its clonal structure. The only values which were not continuous were those between the two stocks of Chrithidia sp. and the remaining stocks of T. cruzi, portraying the difference between the two genera. This continuity of values reflecting an enormous heterogeneity, does not allow typological aproximations to be made in order to describe the genetic variability of T. cruzi. Attempts made by Miles et al. (1977), and Ready and Miles (1980), resulted obsolete as the number of markers and geographical areas analyzed increase. We here entirely agree with the conceptual definition of the clonet, put forward by Tibayrenc and Ayala (1991), above the typical definitions of population or species as the evolutionary unit for this organism.

Spatial autocorrelation analysis - However, two typical clonal characteristics are not apparently observed in the present data, neither in the data from Márquez et al. (1998), with the exeption, maybe, of their clonet 17. We did not found an over-representation of some identical multi-genotype zymodemes in vast areas of Colombia with obvious different enviromental characteristics. Our new Colombian zymodemes are not geographically randomly distributed. As we previously commented, an elevated fraction of the correlograms obtained showed the classic regional patch structure. However, lack of repeated genotypes can be seen in clonal populations when clonal diversity is significant as in the present case, since every stock corresponds to a different genotype.

Otherwise, another clonal feature is not clearly present in our data. There is not a persistence of the same multi-genotypes zymodemes through time in some populations analyzed in different years. It is clearly supported by the dendrograms shown in Fig. 1, where the zymodemes of Ricaurte (Cundinamarca), Tibu and Durania (North of Santander), and Sabanalarga (Casanare) (three different areas of distribution of T. cruzi in Colombia) sampled with some year intervals were not clustered together. In a parallel study, we have investigated some zymodemes of other Colombian stocks of T. cruzi, T. rangeli and some stocks classified as Trypanosoma sp. sampled in different areas, at different times (such as Trypanosoma sp. in Antioquia, Trypanosoma sp. in Ubaqué, Trypanosoma sp. in Arboledas). Fig. 4 shows that the stocks sampled in a same locality through years were not clustered jointly. This is apparently in contradiction with all the characteristics of the clonal population structure discovered in other Latin America areas.

The extraordinary genetic diversity showed at the isoenzymatic level in America by Tibayrenc and collaborators for T. cruzi, together with the 20 new zymodemes found in Colombia alone and presented in this study, could also suggest the possibility of extensive genetic variability of those loci involved in the growth regulation of T. cruzi, its development in specific vectors, as well as those governing diverse pathogenic properties of the organism, which could in turn explain the variety of forms in which Chagas disease manifests itself. This could be so, particularly if we accept the clonal structure, given the possible correlation between groups of variables. A thorough procedure of sampling of populations within the range of distribution of the parasite is therefore of enormous importance to accurately understand the genetic characteristics of T. cruzi in different areas of the American continent.

To collaborative relationships between the Pontificia Universidad Javeriana and the Instituto Nacional de Salud.

Fig. 1 | Fig. 2 | Fig. 3 | Fig. 4 | Table I | Table II | Table III | Table IV | Table V | Table VI | Table VII

This work was partially supported by Colciencias.

  • Alvarez D, Ruiz-Garcia M, Guerrero CJ, Jaramillo-Correa JP 2000. Detection of stabilizant selection in favour of Santa Cruz (SC) homokariotype in Drosophila pseudoobscura populations from the high plateau of Colombian Andes. Gen Mol Biol (in press).
  • Ayala FJ 1975. Genetic differentiation during the speciation process. Evol Biol 8: 1-78.
  • Baum BR, Nevo E, Johnson DA, Beiles A 1997. Genetic diversity in wild barley (Hordeum spontaneum C. Koch) in the Near East: a molecular analysis using Random Amplified Polymorphic DNA (RAPD). Gen Resour Crop Evol 47: 147-157.
  • Bogliolo AR, Lauriapires L, Gibson WC 1996. Polymorphism in Trypanosoma cruzi Evidence of genetic recombination. Acta Trop 61: 31-40.
  • Britton-Davidian J 1990. Genic differentiation in M. m. domesticus populations from Europe, the Middle East and North Africa: geographic pattern and colonization events. Biol J Linn Soc 41: 27-45.
  • Carrasco HJ, Frame IA, Valente SA, Miles MA 1996. Genetic exchange as a possible source of genomic diversity in sylvatic populations of Trypanosoma cruzi Amer J Trop Med Hyg 54: 418-424.
  • Castillo MI, Ruiz-Garcia M 2000. Genetic relationships of the human population of Bogotá (Colombia) with Caucasian, Amerindian, Black and Asian populations by using PM molecular markers. Human Biology (in press).
  • Chakraborty R, Haag M, Ryman N, Stahl G 1982. Hierarchical gene diversity analysis and its application to brown trout population data. Hereditas 97: 17-21.
  • Chapman MD, Caffery A, Miles MA, Swallow DM 1984. Enzyme sub-unit numbers in Trypanosoma cruzi zymodemes. Ann Trop Med Parasit 78: 541-542.
  • Chesser RK 1991a. Gene diversity and female philopatry. Genetics 127: 437-447.
  • Chesser RK 1991b. Influence of gene flow and breeding tactics on gene diversity within populations. Genetics 129: 573-583.
  • Cosenza H, Kroeger A 1991. Enfermedades Parasitarias de Mayor Prevalencia y Transmitidas por Vectores en Centroamérica, Centro de Investigación y Diagnóstico de Enfermedades Parasitarias, 287 pp.
  • Dujardin JP, Tibayrenc M, Venegas E, Maldonado L, Desjeux P, Ayala FJ 1987. Isozyme evidence of lack of speciation between wild and domestic Triatoma infestans (Hemiptera, Reduviidae) in Bolivia. J Med Entomol 24: 40-45.
  • Duran R, Ruiz-Garcia M 2000. Genetic population history of the human population of Bogotá (Colombia) using D1S80, VWA and TH01 molecular markers. Amer J Hum Biol (in press).
  • Dvorak JA 1984. The natural heterogeneity of Trypanosoma cruzi: biological and medical implications. J Cell Biochem 24 : 357-371.
  • Ehrlich PH, Raven PH 1969. Differentiation of populations. Science 165: 228-1232.
  • Ennos RA 1985. The mating system and genetic structure in a perennial grass, Cynosurus cristatus L. Heredity 55: 121-126.
  • Farris JS 1972. Estimating phylogenetic trees from distance matrices. Amer Nat 106: 645-668.
  • Felsenstein J 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791.
  • Hamilton MJ, Kennedy M L 1986. Genic variation in the Coyote, Canis l latrans in Tennessee, USA. Genetica 71: 167-173.
  • Jaramillo-Correa JP, Keyeux G, Ruiz-Garcia M, Rodas C, Bernal J 2000. Population study of the genes APOE, APOB (3' VNTR) and ACE in some black and Amerindian communities from Colombia. Human Heredity (in press).
  • Johnson WE, Selander RK 1971. Protein variation and systematics in kangaroo rats (genus Dipodomys). Syst Zool 20: 377-405.
  • Lewicka K, Breniere-Campana SF, Barbane C, Pedet JP, Tibayrenc M 1995. An isoenzymatic survey of Trypanosoma cruzi genetic variability in sylvatic cycles from French Guiana. Exper Parasitol 81: 20-28.
  • Lopez-Alonso D, Pascual-Reguera L 1989. Population structure and pattern of geographic variation in Muscari comosum along its range of distribution. Genetica 78: 39-49.
  • Márquez E, Arcos-Burgos M, Triana O, Moreno J, Jaramillo N 1998. Clonal population structure of Colombian sylvatic Trypanosoma cruzi J Parasitol 84: 184-190.
  • Miles MA, Toye PJ, Oswald SJ, Godfrey DJ 1977. The identification of isoenzyme patterns of two distinct strain-groups of Trypanosoma cruzi, circulating independently in a rural area of Brazil. Trans R Soc Trop Med Hyg 71: 217-225.
  • Miles MA, Apt WB, Widmer G, Povoa M, Schofield CJ 1984. Isoenzyme heterogeneity and numerical taxonomy of Trypanosoma cruzi stocks from Chile. Trans R Soc Trop Med Hyg 78: 526-535.
  • Nevo E, Beiles A, Kaplan D, Storch N, Zohary D 1986. Genetic diversity and enviromental associations of wild barley, Hordeum spontaneum (Poaceae), in Iran. Pl Syst Evol 153: 141-164.
  • Nevo E, Beiles A, Krugman T 1988. Natural selection of allozyme polymorphisms: a microgeographic climatic differentiation in wild emmer wheat (Triticum dicoccoides). Theor Appl Genet 76: 737-752.
  • Ochman H, Selander RK 1984. Evidence for clonal population structure in Escherichia coli Proc Natl Acad Sci USA 81: 198-201.
  • Ohta T 1982. Linkage disequilibrium due to random genetic drift in finite subdividied populations. Proc Natl Acad Sci USA 79: 1940-1944.
  • Pamilo P 1990. Statistical tests of phenograms based on genetic distances. Evolution 44: 689-697.
  • Randi E, Ragni B 1991. Genetic variability and biochemical systematics of domestic and wild cat populations (Felis silvestris: Felidae). J Mammal 72: 79-88.
  • Ready PD, Miles MA 1980. Delimitation of Trypa-nosoma cruzi zymodemes by numerical taxonomy. Trans R Soc Trop Med Hyg 74: 238-242.
  • Robertson A, Hill WG 1984. Deviations from Hardy-Weinberg proportions, sampling variances and use in estimation of inbreeding coefficients. Genetics 107: 703-718.
  • Ruiz-Garcia M 1993. Analysis of the evolution and genetic diversity within and between Balearic and Iberian cat populations. J Hered 84: 173-180.
  • Ruiz-Garcia M 1994a. Genetic profiles from coat genes of natural Balearic cat populations: an eastern Mediterranean and North African origin. Genet Sel Evol 26: 39-64.
  • Ruiz-Garcia M 1994b. Genetic structure of Marseilles cat population: is there really a strong founder effect? Genet Sel Evol 26: 317-331.
  • Ruiz-Garcia M, Alvarez D 2000. Genetic microstructure in two Spanish cat populations. I: genic diversity, gene flow and selection. Genes and Genetic Systems (in press).
  • Ruiz-Garcia M, Jordana J 1997. Spatial genetic structure of the "Gos d'Atura" dog breed in Catalonia (Spain). Braz J Genet 20: 225-236.
  • Ruiz-Garcia M, Jordana J 2000. Spatial genetic structure from blood allozymes in the Pyrenean Brown, a rare cattle breed, in Catalonia (Spain). Bioch Genetics (in press).
  • Ruiz-Garcia M, Klein K 1997. Genetic structure of populations of the domestic cat in Catalonia (Spain) and upper midwestern USA: a microgeographic and macrogeographic study. J Genet 76: 99-115.
  • Saitou N, Nei M 1987. The neighbor-joining method: a new method for reconstruting phylogenetic trees. Mol Biol Evol 4: 406-425.
  • Saravia NG, Holguin AF, Cibulskis RE, D'alessandro A 1987. Divergent isoenzyme profiles of sylvatic and domiciliary Trypanosoma cruzi in the Eastern Plains, piedmont, and highlands of Colombia. Am J Trop Med Hyg 36: 59-69.
  • Sneath PH, Sokal RR 1973. Numerical Taxonomy, WH Freeman, San Francisco.
  • Sokal RR, Jacquez GM 1991. Testing inferences about microevolutionary processes by means of spatial autocorrelation analysis. Evolution 45: 152-168.
  • Sokal RR, Oden NL 1978a. Spatial autocorrelation in biology. 1. Methodology. Biol J Linn Soc 10: 199-228.
  • Sokal RR, Oden NL 1978b. Spatial autocorrelation in biology. 2. Some biological implications and four applications of evolutionary and ecological interest. Biol J Linn Soc 10: 229-249.
  • Sokal RR, Wartenbeg DE 1983. A test of spatial autocorrelation using an isolation-by-distance model. Genetics 105: 219-237.
  • Sokal RR, Harding R M, Oden NL 1989. Spatial patterns of human gene frequencies in Europe. Amer J Phys Anthrop 80: 267-294.
  • Sokal RR, Oden NL, Barker JSF 1987. Spatial structure in Drosophila buzzati populations: simple and directional spatial autocorrelation. Amer Nat 129: 122-142.
  • Tibayrenc M 1996. Towards a unified evolutionary genetics of microorganisms. Ann Rev Microbiol 50: 401-429.
  • Tibayrenc M, Ayala FJ 1988. Isozyme variability in Trypanosoma cruzi, the agent of Chagas' disease: genetical, taxonomical, and epidemiological significance. Evolution 42: 277-292.
  • Tibayrenc M, Ayala FJ 1991. Towards a population genetics of microorganisms: the clonal theory of parasitic protozoa. Parasitol Today 7: 228-232.
  • Tibayrenc M, Echalar L, Dujardin JP, Poch O, Desjeux P 1985. The microdistribution of Trypanosoma cruzi in southern Bolivia: new isoenzymes profiles and further arguments against Mendelian sexuality. Trans R Soc Trop Med Hyg 78: 519-525.
  • Tibayrenc M, Hoffman A, Poch O, Echalar L, Le Pont F, Lemesre JL, Desjeux P, Ayala FJ 1986a. Additional data on Trypanosoma cruzi isozymic strains encountered in Bolivian domestic transmission cycles. Trans R Soc Trop Med Hyg 80: 442-447.
  • Tibayrenc M, Ward P, Moya A, Ayala FJ 1986b. Natural populations of Trypanosoma cruzi, the agent of Chagas disease, have a complex multiclonal structure. Proc Natl Acad Sci USA 83: 115-119.
  • Weitkamp LR, Arends T, Gallango ML, Neel JV, Schultz J, Shreffler DC 1972. The genetic structure of a tribal population, the Yanomama Indians. III. Seven serum protein systems. Ann Hum Genet 35: 271-279.
  • +
    Corresponding author. Fax: +57-1-285.0503. E-mail:
    Received 8 December 1999
    Accepted 15 August 2000
  • Publication Dates

    • Publication in this collection
      28 Mar 2001
    • Date of issue
      Jan 2001


    • Accepted
      15 Aug 2000
    • Received
      08 Dec 1999
    Instituto Oswaldo Cruz, Ministério da Saúde Av. Brasil, 4365 - Pavilhão Mourisco, Manguinhos, 21040-900 Rio de Janeiro RJ Brazil, Tel.: (55 21) 2562-1222, Fax: (55 21) 2562 1220 - Rio de Janeiro - RJ - Brazil