Molecular Identification Of Rose Variety And Analysis On Genetic Relationship

Plant variety protection (PVP) is a novel system of intellectual property rights protection,which is aimed to protect the right of plant breeders. Some problems were emerged in theprocess of development of PVP, such as rapid identification of origin of plant materials,reducing the cycle of variety test, and molecular biology techniques become the most importanttools in practical operations of PVP. The potential applications of molecular marker techniquesare promising in PVP system. In this study, firstly morphological traits in DUS test wasinvestigated to evaluate phenotypic diversity of rose varieties; secondly SSR and AFLPmarkers were used to identify rose varieties and analyze genetic relationships among thevarieties. A preliminary identification system for rose varieties was established using SSRmarkers. It can be used for rapid identification of propagating materials for infringement cases,moreover it is an ideal approach to construction databases of DNA profiles of commonknowledge and protected varieties. Analysis of genetic relationship could supply some supplyscientific evidence for rose variety breeding and conservation of germplasm resource.1. According to the Rosa guideline for the conduct of test for DUS, the phonotypicaldiversity of Rosa 'Hybrid Tea', Rosa 'Miniature' and Rosa 'Floribunda' was investigated bymeans of field surveys. The result showed that162allele variants were detected in43traits, thenumber of allels in each trait ranged from1to14with an average of3.767; the number ofpolymorphic trait (loci) was40, the ratio of polymorphism is93.02%; the number of effectivealleles per locus ranged from1to7.365with an average of2.358; Shannon information indexranged from0to2.274with an average of0.895and Nei's genetic diversity index ranged from0to0.864with an average of0.487. The cluster analysis and principal component analysisindicated that genetic variation was evident among different horticultural rose groups, but afew varieties were overlapped among different horticultural groups.2. Analysis of polymorphism in SSR loci showed that460allele variants were detected in41SSR loci, the number of alleles at each locus ranged from2to19with an average of11.2; the average number of allele variants was13.6at17genomic SSR loci, and9.5at24EST-SSRloci; the number of effective alleles ranged from1.994to8.286at41loci, the greatest numberwas found at locus Rw55E12; the number of allelic phenotypes range from3to92, the valueof polymorphic information content ranged from0.498~0.864. It indicated that thepolymorphism of SSR locus was reliable, and genomic SSR was superior to EST-SSR.3. Molecular identification of rose varieties based on SSR marker showed that thediscriminating power was flexible at41SSR loci; the number of unique allelic phenotypesrange from0to67, and the largest was at Rw10M24, the least was at H1F03, the value of Djranged from0.315~0.988, the loci with largest values were CTG623and Rw55E12, the leastwas H1F03; three pairs of rose varieties shared the same allelic phenotype at41SSR loci,respectively; three SSR loci were necessary to discriminate the rest rose varieties completely atleast, such as combination of RA013a,RA043a and Rw5D11, combination of P1,Rw8B8andRw59A12, and combination of Rw10M24,Rw5D11and CTG21. It was found that the SSRDNA profiles were identical between the original rose variety and its mutatant varieties, andthe divergence of SSR DNA profiles among hybrid varieties was obvious. It was suggested thatSSR molecular marker are feasible in rapid identification of plant varieties.4. The result of analysis of different horticultural rose groups showed that all sampleswere divided into six groups. The first group included13varieties of Rosa 'Chinas' and a fewones of Rosa 'Shrub'; the second group was mainly made up of the varieties of Rosa 'Hybridtea'; the third group consisted of a few varieties of Rosa 'Hybrid Tea', Rosa 'Floribunda', Rosa'Shrub' and Rosa 'Climber'; the forth group included those of Rosa 'Polyantha' and some ofRosa 'Shrub'; the fifth group consisted of most varieties of Rosa 'Shrub'; the sixth groupincluded those of Rosa 'Mininature' and Rosa 'Hybrid Tea'. It was suggested that commonorigin and related rose varieties be clustered together based on SSR molecular markers. Theresult of clustering and principal components analysis based on molecular markers was similarto horticultural group in part. The result of correlation analysis between morphological dataand SSR data showed weak correlation, with a correlation coefficient of0.321. 5. The result of variety identification and analysis of genetic relationship showed that1803DNA fragments were amplified in14primer combinations,1727fragments werepolymorphic, and percentage of polymorphic bands was93.02%; pairwise Dice dissimilarityin rose varieties ranged from0.111to0.631, it suggested that the genetic dissimilarity based onAFLP markers can be used for variety identification. The genetic dissimilarity of full-sibsfamily variety and related variety was small, which confirmed that AFLP is an effective tool forvariety identification. The result of clustering and principal components analysis based onAFLP markers was similar to the result on SSR markers; genetic divergence betweenhorticultural groups was distinct, but genetic differences in horticultural group was present.Correlation analysis indicated that there was a weak correlation between AFLP marker data andSSR marker data, with a correlation coefficient of0.323.6. Comparison of different genetic similarity coefficients based on microsatellite markersshowed, pairwise correlation coefficient ranged from0.735to1.000; cophenetic correlationcoefficient ranged from0.835to0.923; consensus fork index (CIc) between differentdendrograms which were constructed by different genetic similarity coefficients ranged from0.404to1.000, it indicated that the result of clustering varied according to different coefficients.The value of STRESS ranged from17.19~27.92%, Russell and Rao coefficient, Dicecoefficient, Jaccard coefficient and Ochiai coefficient were at the same level. Considering suchcharacteristics as molecular markers, analysis of goodness of fit, algorithm of coefficients, andcomparison between result of clustering and lineage analysis, it was suggested that Dicecoefficient, Jaccard coefficient and Ochiai coefficient are most appropriate for genetic analysisbased on microsatellite data from roses, and the next appropriate was the Simple Matchingcoefficient. Russell and Rao coefficient, Yule coefficient, Hamann coefficient and Phicoefficient should be avoided as much as possible.