Genetic base of Brazilian irrigated rice cultivars

The aim of this study was to estimate the genetic base of Brazilian irrigated rice cultivars released in the period from 1965 to 2012. The genealogies of the cultivars were obtained based on information from marketing folders, websites, crossings records, and scientific articles. The following factors were calculated: relative genetic contribution (RGC), accumulated genetic contribution (AGC), frequency (in percentage) of each ancestor in the genealogy (FAG), number of ancestors that constitute each cultivar (NAC), number of ancestors responsible for 60%, 70%, 80% and 90% of the genetic base (NAGB), and average number of ancestor per cultivar (ANAC). The cultivars were also grouped based on the period of release (1965-1980, 1981-1990, 1991-2000 and 2001-2012). For each grouping, the previously described factors were also estimated. A total of 110 cultivars were studied and it was concluded that the genetic base of Brazilian irrigated rice cultivars is narrow.


INTRODUCTION
Rice is considered to be the cereal crop of greatest socioeconomic importance since it is one of the staple food components and the main dietary energy source for the human population.Brazil is the eighth largest producer of this grain, and Brazilian production in the 2010/2011crop season was estimated at 9.45 million tons (CEPA 2011).The high productivity rates obtained in the country are mostly due to the activity of research institutes, mainly through development of new cultivars and crop techniques.
The first rice genetic breeding activities in Brazil began in 1936, at the Instituto Agronômico de Campinas (IAC), when the first competitive tests were carried out.After that, in 1938, the first hybridizations were carried out between varieties introduced from other countries and local breeds (Viegas et al. 1945).In the same year, the Instituto Rio Grandense de Arroz (IRGA) began its irrigated rice breeding program (Soares et al. 2004a).Up to the 1970s, rice breeding in Brazil was carried out basically through the activities of IAC and IRGA.As of 1970, other programs were established at federal and state level.The development of domestic cultivars spurred Brazilian rice growing, making it competitive and profitable in view of yield increases in the order of 30% registered in the middle of the 1980s (Morais and Rangel 1997).
The Brazilian irrigated rice cultivars were developed from lines and cultivars coming from international research institutes, such as the International Rice Research Institute (IRRI) in the Philippines (Morais et al. 2004), the International Center for Tropical Agriculture (CIAT), and the Institute for Research in Tropical Agriculture (IRAT), currently called the Centre de Coopération Internationale en Recherche Agronomique pour le Développement (CIRAD), headquartered in France.These genotypes were introduced in Brazil by domestic breeding programs, and they represent most of the genetic base of Brazilian irrigated rice cultivars.The genetic base is defined by Cui et al. (2000) as the entire set of genes that contribute toward the development of cultivars of a determined species.
Since they originate from a restricted group of ancestors, from crosses with kinship relations, and for having been developed by breeding methods that do not allow broad recombination, such as genealogical, populational and backcrossing methods (Cordeiro 2008), Brazilian cultivars have a narrow genetic base (Rangel et al. 1996).
Narrowing of the genetic base is considered one of the main drawbacks of modern agriculture since it causes genetic vulnerability of cultivars to biotic and abiotic stresses, which may result in yield reduction in the rice crop (Rangel et al. 1996).For genetic breeding purposes, narrow genetic base results in low variability, which may limit possibilities for recombination and genetic segregation, and hinder achievement of gains from selection.Knowledge regarding the magnitude of the genetic base of Brazilian irrigated rice cultivars may help in the choice of divergent parents in breeding programs, contributing to the broadening of the genetic base of cultivars and cultivated species.
In order to provide assistance to genetic breeding programs, the aim of this study was to estimate the genetic base of Brazilian irrigated rice cultivars released in the commercial market in Brazil in the period from 1965 to 2012, except for the cultivars developed by private companies, since their genealogical records are not disclosed.

MATERIAL AND METHODS
Irrigated rice cultivars released for commercial use in the period from 1965 to 2012 were studied.The purpose was to construct the genealogy of each cultivar for the identification of divergent ancestors.During the preparation of the genealogies, it was considered information obtained from marketing folders of varieties, websites, genealogical records of research institutes, such as the IRRI (1985IRRI ( , 1998) ) and CIAT (2005), and scientific articles (Dilday 1990, Rangel et al. 1996, Montalván et al. 1998, Silva et al. 1999, Châteal and Guimarães 2002, Cutrim and Rangel 2002, Soares et al. 2005, Rangel et al. 2007, Cordeiro and Medeiros 2010a).Hybrid cultivars and a large part of the cultivars developed by private companies were not evaluated since their genealogical records are not disclosed.
The relative genetic contribution (RGC) of each ancestor was estimated considering that there is no degree of kinship between the ancestors, and that the theoretical proportion of genes transferred from a parent to its offspring is 50% (Delannay et al. 1983).This is estimated from the sum of the individual contributions of each ancestor in all the cultivars, according to the methodology used by Rangel et al. (1996), Montalván et al. (1998) and Silva et al. (1999).The ancestors were classified in decreasing order of the RGC values, and the Accumulated Genetic Contribution (AGC) was estimated from the successive sum of the RGC of each ancestor.
The frequency (in percentage) of each ancestor in the genealogy (FAG) was estimated considering the number of cultivars that had a certain ancestor in their genealogy in relation to the total number of cultivars analyzed.For each cultivar, it was estimated the number of ancestors that constitute them (NAC).
The cultivars were also grouped according to the year of commercial release, establishing four periods delimited by the years 1965-1980, 1981-1990, 1991-2000 and 2001-2012.For each grouping it was determined: number of ancestors responsible for 60%, 70%, 80% and 90% of the genetic base (NAGB), number of ancestors and cultivar (NA and NC, respectively), ratio between the number of ancestors and number of cultivars (NA/NC), and average number of ancestors per cultivar (ANAC).The latter corresponds to the sum of each NAC divided by the total number of cultivars.

RESULTS AND DISCUSSION
The genealogical information from 110 Brazilian irrigated rice cultivars was obtained, which were developed from 123 different ancestors (Table 1  The greatest NAC are attributed to BRS Pampa (NAC= 41) and BRSMG Seleta (NAC= 39).Both of them have two sources of resistance to rice blast (Pyricularia grisea) coming from the ancestors Tetep and Tadukan (Rangel et al. 1996).According to Magalhães Júnior (2011), BRS Pampa has intermediate to moderate resistance to rice blast and according to Soares et al. (2004b), BRSMG Seleta is resistant.
The 17 cultivars released commercially in the period from 1965 to 1980 have a narrow genetic base, and they were developed from 32 ancestors (Table 3), with eight cultivars being introduced from other countries (Belle Patna, Bluebelle, Bluebonnet, Bluebonnet 50, Cica 4, Cica 7, Dawn and Le Bonnet).Eight of these ancestors represented approximately 70% of the genealogy, and Marong Paroc, Dee Geo Woo Gen, Cina and Latisail ancestors together contributed to 45% of the genealogy (Table 4).The narrowing of the genetic base in this period from 1965 to 1980 is more accentuated, for around 90% of the genes are represented by 15 ancestors, and the cultivars consist of an average of 1.88 ancestors (Table 3).
In the period of 1981-1990, twenty-five new ancestors were incorporated in the genetic base of the Brazilian irrigated rice cultivars.In numerical terms, there was no alteration of the genetic base, since approximately 70% of the gene pool originate from eight ancestors, just as in the previous period (Table 3).In the period of 1991 -2000 it began what was considered as an increase in the genetic base, since around 70% of the genealogy of the 36 cultivars released commercially in this period originate from 11 ancestors (Table 3).Nevertheless, just as in the previous period, a preference was seen for the use of more adapted genotypes in the development of the cultivars, since Dee Geo Woo Gen, Cina, Latisail and MCVA are responsible for around 50% of the gene pool.The genetic base with the greatest number of ancestors was observed for the 26 cultivars released commercially from 2001 to 2012.In this period, ninety-two ancestors were used, thirty-nine of them being new, and approximately 60% of the genetic base is represented by 14 ancestors, practically double the average of the previous periods (Table 3).The greatest values of ANAC (20.19) and NA/NC (3.53) are also attributed to this period, which highlight the cultivars broad genetic base.
In spite of the large number of ancestors, the genetic base of Brazilian upland rice cultivars is considered to be narrow, since around 70% of the gene pool studied is derived from 13 ancestors.Cuevas-Pérez et al. (1992) obtained similar results for irrigated rice cultivars recommended from 1971 to 1989 for the regions of Latin America and the Caribbean.Similar to the present study, it was observed that 14 ancestors represented 70% of the genetic base, and that Cina, Latisail and Dee Geo Woo Gen were the most used ancestors.Montalván et al. (1998) report a similar narrowing for upland rice cultivars released for commercial use from 1971 to 1993; it was observed that approximately 70% of the genetic base is derived from the Brazilian ancestors Dourado Precoce, Pérola and Pratão.
When ancestors with RGC of less than 2% are disregarded, the genetic base becomes even narrower, being restricted to ten ancestors, which represent 66% of the entire gene pool (Table 1).Rangel et al. (1996) report greater narrowing of the genetic base of irrigated rice cultivars recommended for Brazil in the 1992/93 crop season, also obtaining an effective size (N e ) of ten ancestors, when considering only those with RGC above 2%.The same ten ancestors contributed to 68% of the genes of the cultivars used in the period from 1980 to1992.
The average number of ancestors per cultivar (ANAC) was estimated at 14.37 (Table 3).This parameter should be used with restriction in the study of genetic base, since it does not represent the amplitude of contributions from parents for the cultivars.In addition, within each cultivar, the genetic contributions of the ancestors vary in magnitude, such as the cultivar IRGA 425, which, in spite of being Table 3. Number of cultivars (NC), number of ancestors (NA), average number of ancestor per cultivar (ANAC), number of ancestors responsible for 60%, 70%, 80% and 90% of the genetic base (NAGB), number of new ancestors added to the genetic base for each reporting period and ratio between the number of ancestors and the number of cultivars (NA/NC) of Brazilian irrigated rice cultivars released on the commercial market in the period from 1965  The ratio between the number of ancestors and the number of cultivars (NA/NC) was 1.12, and it is considered low, comparing with values of other studies in which the genetic base was considered narrow.Cuevas-Pérez et al. (1992) obtained a value of 1.25 for the irrigated rice cultivars recommended for Latin America and the Caribbean from 1971 to 1989, and Montalván et al. (1998) obtained a value of 1.29 for Brazilian upland rice cultivars.Therefore, the value of NA/NC obtained in this study may be one more indication of the narrow genetic base of Brazilian irrigated rice.
The low value of NA/NC is brought about by the use of the same ancestors in the composition of various cultivars due to the preference of breeding programs in using the most adapted genotypes as parents in different crossings.
In relation to the use of genotypes from breeding programs in crossings, Nass and Parteniani (2000) emphasize that wild accessions, or those that are not a result of breeding programs, have been left aside by breeders since the use of new sources of variability requires prior work for characterization and more selection cycles for obtaining superior genotypes.
The ancestor Dee Geo Woo Gen is involved in the formation of 89.19% of the cultivars evaluated.According to Soares et al. (2004a), this genotype, as well I Geo Tze, is Chinese cultivar that bear a recessive allele which confers low height, and, exclusively in these cultivars, the dwarf allele does not affect the stretching out of spikelets and panicles.Therefore, the use of these ancestors seeks to obtain semi-dwarf size cultivars, a trait that is desirable for modern rice cultivars.
The ancestors Cina and Latisail were likewise used in 83.78% of the cultivars.The high frequency of these genotypes in the genealogy is due in part to the fact of being the genitors of Peta cultivar -which, for its part, is the female parent of IR 8, with Dee Geo Woo Gen being the male parent.Released by the IRRI in 1966, IR 8 has semi-dwarf height, insensitivity to photoperiod, high tillering and high yield.IR 8 is considered as the first cultivar of the Indica group, which is highly productive and adapted to tropical climates (Soares et al. 2004a).Since it aggregates various agronomic traits favorable to the climatic conditions of Brazil, IR 8 was widely used in the development of Brazilian cultivars.
The narrowing of the genetic base in the period from 1965 to 1980 is more accentuated, for around 90% of the genes are represented by 15 ancestors, and the cultivars consist of an average of 1.88 ancestors (Table 3).In the period from 1981 to 1990, 70% of the gene pool came from eight ancestors, just like on the previous period.However, it can be considered that there was genetic narrowing, since the ancestors Dee Geo Woo Gen, Latisail, Cina and MCVA represented 52% of the total genealogy, and Dee Geo Woo Gen was the one that stood out (Table 4) All the cultivars released commercially in the period from 1991 to 2000 have some degree of kinship, since 100% of them have genes from Dee Geo Woo Gen (Table 4).In the period from 2001 to 2012 it was observed some widening in genetic base for the 26 cultivars released Table 4. Relative genetic contribution (RGC%), accumulated genetic contribution (AGC%) and frequency of the ancestors in the genealogy (FAG%) of the principal ancestors of Brazilian cultivars of rice released on the commercial market in the four reporting periods (1965-1980, 1981-1990, 1991-2000, 2001- commercially.In this period, 92 ancestors were used, 39 of them being new and approximately 60% of the genetic base is represented by 14 ancestors, practically double the average of the previous periods (Table 3).This may be due to disclosure of the genealogies and studies on the genetic base of Brazilian cultivars in the 1990's.
The greatest values of ANAC (20.19) and NA/NC (3.53) are also attributed to this period, which highlight the broad genetic base of the cultivars.Nevertheless, the genetic base may be considered as unbalanced, for only 38 ancestors represent around 90% of the genealogy, and the remaining 50 contribute to less than 10%.Just as in most of the previous periods, Dee Geo Woo Gen, Cina and Latisail were the most used ancestors.This practice restricts the possibilities for recombination in breeding programs and increases the genetic uniformity of the nucleus and cytoplasm, which may result in vulnerability of the cultivars to biotic and abiotic stresses and restriction of gains from selection, especially for quantitative traits (Vello et al. 1988, Rangel et al. 1996, 2000).
Aiming at expanding the genetic base, Rangel et al. (1996) suggest some alternatives.One of them would be the crossing of divergent parents originating from other breeding programs for the desired agronomic traits.Another option would be the use of multiple crossings between accessions of exotic germplasm and non-kin elite genotypes, or genotypes with a small degree of kinship.A third strategy consists of using recurrent selection for composition of populations with a broad genetic base; nevertheless, the high cost of implementation of this method impedes its adoption in breeding programs (Canci et al. 1997).
The use of male-sterile lines may make the use of recurrent selection viable, since it reduces operational costs, and favors natural recombination (Coimbra et al. 2008).However, the use of male-sterility is restricted due to the complexity of handling and maintenance of the genotypes bearing the sterility genes.Rangel et al. (1996) also indicate the use of wild rice species as another alternative for expansion of the genetic base.In this context, the use of O. glumaepatula in hybridizations with O. sativa is an efficient alternative for increasing the genetic base of rice, as well as providing the attainment of genotypes with high yield and desirable agronomic traits, which are able to be used as parents in new crosses (Cordeiro and Medeiros 2010b).
It may be seen that the narrow total genetic base of Brazilian irrigated rice cultivars is due to the fact that the genetic breeding programs for Brazilian rice began from a reduced number of genotypes, nearly all of them coming from other countries.Over the years, there was incorporation of new sources of variability, which contributed to expansion of the genetic base.This may have arisen from awareness on the part of breeders in regard to the risks of genetic base narrowing.Nevertheless, the use of new genotypes appraisable by genetic breeding and better utilization of currently available genetic resources are necessary since many of the ancestors of the genetic base of Brazilian irrigated rice are little used.

CONCLUSIONS
The genetic base of Brazilian irrigated rice cultivars in the period from 1965 to 2012 is narrow.
The narrowing of the genetic base is mainly brought about by the use of the same genotypes as parents of various crossing.
The genetic base of cultivars released commercially in the period from 2001 to 2012 is the broadest among the periods evaluated.Nevertheless, many of the ancestors used in this period contribute in small proportion to the genetic base.

Table 1 .
Relative genetic contribution (RGC%), accumulated genetic contribution (AGC%) and frequency of the ancestors in the genealogy (FAG%) of Brazilian cultivars of rice released on the commercial market in the period from 1965 to 2012 (Magalhães Júnior et al. 2010)2-9 Formosa and BR 3 Caeté were not included Table2).Cultivars Bluebonnet, Bluebonnet 50 and BRS Sinuelo CL presented the lowest number of ancestors per cultivar (NAC= 2); Bluebonnet and Bluebonnet 50 are derived from the crossing between Rexoro and Fortuna cultivars.BRS Sinuelo CL was obtained by three cycles of backcrossing between BRS7 Taim and the genotype AS 3510; the latter contains a gene, which confers tolerance to imidazolinone-based herbicides used mainly for red rice control in commercial crops(Magalhães Júnior et al. 2010).Nevertheless, since the parents of BRS 7 Taim are not known, it is possible that BRS Sinuelo CL has a broader genetic base than it has AncestorsRGC (%) AGC (%) FAG (%) Ancestors RGC (%) AGC (%) FAG (%)

Table 2 .
Number of ancestors per cultivar (NAC) of 110 Brazilian irrigated rice cultivars released on the commercial market in the period from 1965 to 2012 2012)