选用204对SSR标记对全国野生大豆群体(174份代表性样本)的基因组扫描,采用TASSEL软件的GLM(general linear model)方法对百粒重、开花期、成熟期、干豆腐得率、干豆乳得率和耐淹性性状值关联分析,解析与性状关联位点的优异等位变异,鉴别出一批与农艺、加工性状关联的优异等位变异及携带优异等位变异的载体材料;进一步分析极值表型材料的遗传构成。结果表明:(1)累计51个位点(次)与性状关联,有些标记同时与2个或多个性状相关联,可能是性状相关的遗传基础;关联位点中累计16位点(次)与连锁分析定位的QTL一致;(2)与地方品种群体和育成品种群体的关联位点比较,发现野生群体关联位点只有少数与之相同,群体间育种性状的遗传结构有明显差异。(3)与多性状关联的位点其等位变异对不同性状的效应方向可相同可不同,如GMES5532a-A332对百粒重和耐淹性的相对死苗率都是增效效应,而GMES5532a-A344对百粒重是减效效应,对相对死苗率是增效效应;(4)极值表型材料间的遗传构成有很大差异。表型值大的材料携带较多增效效应大的位点等位变异,例如N23349的百粒重是9.08g,含有4个增效效应较大的位点等位变异;表型值小的材料携带较多减效效应大的位点等位变异,如N23387的百粒重是0.75g,含有4个减效效应较大的位点等位变异。关联作图得到的信息可以弥补连锁定位信息的不足,尤其是全基因组位点上复等位变异的信息为育种提供了亲本选配和后代等位条带辅助选择的依据。
This study was designed to reveal the genome‐wide distribution of presence/absence variation（PAV） and to establish a database of polymorphic PAV markers in soybean. The 33 soybean whole‐genome sequences were compared to each other with that of Williams 82 as a reference genome. A total of 33,127 PAVs were detected and 28,912 PAV markers with their primer sequences were designed as the database NJAUSoyPAV_1.0. The PAVs scattered on whole genome while only 518（1.8%） overlapped with simple sequence repeats（SSRs） in BARCSOYSSR_1.0database. In a random sample of 800 PAVs, 713（89.13%） showed polymorphism among the 12 differential genotypes. Using 126 PAVs and 108 SSRs to test a Chinese soybean germplasm collection composed of 828 Glycine soja Sieb. et Zucc. and Glycine max（L.） Merr. accessions, the per locus allele number and its variation appeared less in PAVs than in SSRs. The distinctness among alleles/bands of PCR（polymerase chain reaction） products showed better in PAVs than in SSRs, potential in accurate marker‐assisted allele selection. The association mapping results showed SSR t PAV was more powerful than any single marker systems.The NJAUSoyPAV_1.0 database has enriched the source of PCR markers, and may fit the materials with a range of per locus allele numbers, if jointly used with SSR markers.
