應用雙限制酶切位點標定法進行臺灣藜種原遺傳歧異度分析

Abstract

臺灣藜 (Chenopodium formosanum Koidz.) 為臺東縣重要特色雜糧作物，現有栽培種原多為原住民族部落流傳下來之地方品系，因年代久遠，品系多混雜，且相關之遺傳與育種等研究甚少。為了解種原遺傳背景及提升品種改良效率，本研究以臺東區農業改良場自2009年於臺東縣、花蓮縣及屏東縣等地蒐集的39個種原為材料，為探討臺灣藜與其他藜屬種原之關係，再加入藜麥 (Chenopodium quinoa Willd.) 種原及Chenopodium berlandieri nuttaliae種原共2個種原，總計41個種原作為試驗材料。研究方法使用雙限制酶切位點標定法 (double digest Restriction Associated DNA sequencing, ddRAD) 建立定序圖書庫，次世代定序 (Next generation sequencing, NGS) 採用Illumina HiSeq平台，定序結果以Stacks軟體進行分析，篩選出1,046個具多型性的基因座供後續族群結構、主成分分析及集群分析。種原代號98T011因定序量不足，排除於後續的分析。族群結構分析以STRUCTURE軟體進行，40個種原在K=2時，ΔK呈最高峰值，顯示K=2為最佳次族群分群數；臺灣藜種原在K=3時，ΔK呈最高峰值，顯示K=3為最佳次族群分群數。主成分分析及集群分析以R軟體進行分析，40個種原分析結果，將臺灣藜種原與其他藜屬種原 (藜麥及C. berlandieri nuttaliae) 分為兩群；臺灣藜種原分析結果則依蒐集地區分為兩大群，株型直立、黑色籽實之98T020種原的遺傳背景與其他種原差異較大，另成一群，本結果明確臺灣藜與國外藜麥之親緣關係，並可提供未來臺灣藜育種利用與種原保存策略之參考。

Djulis (Chenopodium formosanum Koidz.) is an important specialty cereal crop in Taitung. Current cultivars are almost local lines conserved by aboriginal tribes. Due to the traditional djulis cultivation habit is remote from now, local lines are mixed by multiple lines, and the related genetics and breeding studies are also deficient. In order to understand the genetic background among djulis germplasms and improve the breeding efficiency, this study used 39 djulis accessions collected in Taitung, Hualien, and Pingtung by Taitung District Agricultural Research and Extension Station in 2009. And 2 Chenopodium genus but different species accessions from NPGRC. In total 41 accessions as the material, to understand the relationship between djulis and other Chenopodium genus. We used double-digest RAD sequencing to build the library, and Illumina HiSeq platform for Next generation sequencing. Data was analyzed by Stacks software, resulting 1,046 polymorphic loci for genetic structure analysis, principal coordinate anlysis and cluster analysis. The sequencing quantity of 98T011 accession was not enough and was excluded from the following analysis. Genetic structure analysis was analyzed by STRUCTURE software. When K=2, 40 accessions had the maximum ΔK, indicated that K=2 was the best cluster number of subpopulation. When K=3, djulis accessions had the maximum ΔK, indicated that K=3 was the best cluster number of subpopulation. Principal coordinate anlysis and cluster analysis were analyzed by R software, the result of the 40 accessions showed that the djulis and the accessions from NPGRC were clearly divided into two groups. As for the result within the djulis accessions, djulis accessions were divided into two groups according to the region they had been collected. The genetic background of the accessions 98T020 was different from other accessions, and was classified to another groups. The result of this study confirm the relationship between djulis and quinoa, and providing the strategy for djulis breeding and germplasm conservation.