利用PCR-RFLP技術鑑定‘玉珍’與‘豐富二號’蘿蔔自交不親和基因，比較SLG序列變異和親緣關係，進一步預測其自交不親和指數。抽取兩品種蘿蔔的DNA作為PCR反應的DNA模板，經PCR擴增得到大約1160 bp片段。共計使用12限制酶，其中以Sal I、Sty I、Hae Ⅲ和Xho I等限制酶能切出3-7片段最適合使用。利用限制酶圖譜和基因定序方法共發現12 SLGs基因。由‘豐富二號’和‘玉珍’蘿蔔分離得到SLGs基因序列演譯成蛋白質序列後，利用NetNGlyc 1.0軟體分析發現SLG3B蛋白質有N-glycosylation位置最多有9個，SLG2B最少只有4個。基因序列與蛋白質序列之親緣樹狀圖發現12個蘿蔔SLGs平均分散於樹狀圖中。‘豐富二號’蘿蔔的SLG1A，SLG2A，SLG4A與SLG5B，SLG6A兩群的親緣樹關係最近，花旗親和指數除了SLG1A和SLG4A分別為2.00和2.50外，其餘皆為0；SLG2B與SLG3A成一個群，花期親和指數分別為2.00和1.00，根據SLG序列的基因和蛋白質核酸序列之親緣關係，可以推測花期親和性指數，但精確的推測須增加資料庫量才能進一步預測。

In the experiment, the identification of SLG-alleles of radish was based on the PCR-RFLP method; further the SLG-genes were sequenced and proteins were to deduced construct phylogenetic tree that predicted the incompatibilities of S-alleles of radish. The genomic DNAs extracted from radish cvs. Rabano and Feng-fu No.2 were used as the template for PCR reaction. The amplified DNA fragments having length about 1160 bp were digested by 12 restriction enzymes, these fragments were digested into 3-7 subfragments by Sal I, StyI, Hae Ⅲ, Xho I. in the experiment the 12 SLGs were found based on the restriction enzyme patterns and gene sequences. The conserved N-glycosylation sites of the 12 SLG proteins deduced from gene sequences were predicted with the NetNGlyc 1.0 program, the results showed that the SLG3B has the highest number of 9 sites and the SLG2B only has the lowest of 4 sites. Gene and protein sequences of the 12 S-locus glycoproteins and other deposits of radish, cabbage and chinese cabbage in GeneBank, NCBI were pooled to construct the phylogenetic tree. The results showed that both cluster of SLG1A, SLG2A, SLG4A and SLG5B, SLG6A appeared the closest relationship of phylogenetic tree, the idex of compatibility of open flower pollination were 0 except SLG1A and SLG4A. Proteins both SLG2B and SLG3A were in same cluster they showed high idex of compatibility of open flower pollination. It was concluded that S-alleles incompatibilities might be predicted based on the phylogenic tree of sequences of genes or proteins, it may be improved by scale up databank.