Abstract: 【Objective】 The present study was aimed to take a first step of the improvement of the previously reported chromosome segment substitution line (CSSL) population SojaCSSLP1, and to explore superior QTL/gene-alleles related to some agronomic traits from the wild parent (Glycine soja Sieb. et Zucc.) for broadening the genetic basis of cultivated soybean (Glycine max (L.) Merr.). 【Method】 The SojaCSSLP1, with the wild soybean N24852 as donor parent and the cultivated soybean NN1138-2 as recurrent parent, was treated with adding markers and removing a number of lines with segment of single marker, the new population was designated as SojaCSSLP2. By using the new population, the QTL/segments for flowering time (FT), plant height (PH), node number (NN), pod number per plant (PN), 100-seed weight (100SW) and seed weight per plant (PSW) were detected through joint comparisons among CSSLs significantly different from the recurrent parent based on QTL mapping with the methods of single marker analysis (SMA), interval mapping (IM), inclusive composite interval mapping (ICIM) and mixed linear composite interval mapping (MCIM), for experiments in three years each with two locations. 【Result】SojaCSSLP2 was composed of 150 CSSLs, of which 130 ones were the same as SojaCSSLP1, added 40 new SSR markers into the previous molecular map, resulted in the average genetic distance and the number of lines with genetic distance more than 30 cM between adjacent markers reduced from 16.15 cM and 32 to 12.91 cM and 17, respectively. The total length of the genetic map increased by 103.52 cM in comparison to the original map (2 063.04 cM). The genome component of NN1138-2 in CSSLs ranged from 79.45% to 99.70% with an average of 94.62% in SojaCSSLP2. Tested in three years and two locations, 4, 5, 5, 7, 14 and 3 working QTL/segments were identified for FT, PH, NN, PN, 100SW and PSW, respectively, with the improved population. Of those, 15 QTL/segments were joint working QTL which could be detected in more than one environment. The directions of additive effects for wild alleles were consistent with that expected from the parental phenotypes except segment of Sct_190-Sat_293 for NN. Among the QTL, the individual QTL could explain 5% to 64% of the phenotypic variation. There were 3, 2 and 2 QTL/fragments that interacted with locations for PH, NN and NP, respectively, and the interactions associated with Fengyang could increase the value of phenotype, which may be due to its higher latitude than that of Nanjing. These detected loci distributed on 26 substituted segments, among them seven related to more than one trait, which might be the genetic basis of correlation among the traits. Compared with the results in the literature, 3, 3, 2, 2, 8 and 2 QTL could also be detected in other cultivated soybeans for FT, PH, NN, 100SW and PSW, respectively, indicating allele differentiation happened not only between wild and cultivated but also among cultivated soybeans. The other 18 loci/segments were newly discovered in the wild soybean.【Conclusion】 The genetic base of FT, PH and NN are much simpler than that of 100SW. There was a large effect QTL (PV＜10%) in the former, while the later is controlled by many small effect loci (PV＜10%) with complicated genetic bases. Using the wild soybean, the novel alleles with the capability of broadening the genetic base of cultivated soybean can be explored.

Key words: wild soybean (Glycine soja Sieb. et Zucc.), cultivated soybean (Glycine max (L.) Merr.), chromosome segment substitution line (CSSL), agronomic trait